Last active
May 6, 2016 11:01
-
-
Save stas-sl/6a47eab2ef88d8fc5a201242689fa9c9 to your computer and use it in GitHub Desktop.
This file has been truncated, but you can view the full file.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| I0506 00:21:17.931069 15760 solver.cpp:280] Solving mixed_lstm | |
| I0506 00:21:17.931088 15760 solver.cpp:281] Learning Rate Policy: fixed | |
| I0506 00:21:18.209065 15760 solver.cpp:229] Iteration 0, loss = 27.5695 | |
| I0506 00:21:18.209156 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:21:18.209175 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:21:18.209188 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:21:18.209200 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:21:18.209211 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:21:18.209223 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:21:18.209236 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0 | |
| I0506 00:21:18.209247 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.125 | |
| I0506 00:21:18.209264 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0 | |
| I0506 00:21:18.209276 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0 | |
| I0506 00:21:18.209288 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0 | |
| I0506 00:21:18.209300 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0 | |
| I0506 00:21:18.209311 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0 | |
| I0506 00:21:18.209322 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0 | |
| I0506 00:21:18.209333 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0 | |
| I0506 00:21:18.209352 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0 | |
| I0506 00:21:18.209372 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0 | |
| I0506 00:21:18.209391 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0 | |
| I0506 00:21:18.209410 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0 | |
| I0506 00:21:18.209422 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0 | |
| I0506 00:21:18.209434 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 0 | |
| I0506 00:21:18.209445 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 0 | |
| I0506 00:21:18.209456 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 0 | |
| I0506 00:21:18.209467 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0 | |
| I0506 00:21:18.209487 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0 | |
| I0506 00:21:18.209506 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.31111 (* 0.3 = 1.29333 loss) | |
| I0506 00:21:18.209522 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 4.35963 (* 0.3 = 1.30789 loss) | |
| I0506 00:21:18.209568 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.27506 (* 0.0272727 = 0.116593 loss) | |
| I0506 00:21:18.209583 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.31833 (* 0.0272727 = 0.117773 loss) | |
| I0506 00:21:18.209596 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.30486 (* 0.0272727 = 0.117405 loss) | |
| I0506 00:21:18.209610 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 4.31247 (* 0.0272727 = 0.117613 loss) | |
| I0506 00:21:18.209625 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 4.26995 (* 0.0272727 = 0.116453 loss) | |
| I0506 00:21:18.209638 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 4.32896 (* 0.0272727 = 0.118062 loss) | |
| I0506 00:21:18.209652 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 4.26778 (* 0.0272727 = 0.116394 loss) | |
| I0506 00:21:18.209667 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 4.31872 (* 0.0272727 = 0.117783 loss) | |
| I0506 00:21:18.209686 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 4.26878 (* 0.0272727 = 0.116421 loss) | |
| I0506 00:21:18.209707 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 4.35593 (* 0.0272727 = 0.118798 loss) | |
| I0506 00:21:18.209729 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 4.37412 (* 0.0272727 = 0.119294 loss) | |
| I0506 00:21:18.209750 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 4.31432 (* 0.0272727 = 0.117663 loss) | |
| I0506 00:21:18.209764 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 4.30842 (* 0.0272727 = 0.117502 loss) | |
| I0506 00:21:18.209777 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 4.34458 (* 0.0272727 = 0.118489 loss) | |
| I0506 00:21:18.209791 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 4.28856 (* 0.0272727 = 0.116961 loss) | |
| I0506 00:21:18.209815 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 4.3757 (* 0.0272727 = 0.119337 loss) | |
| I0506 00:21:18.209830 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 4.34243 (* 0.0272727 = 0.11843 loss) | |
| I0506 00:21:18.209846 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 4.31728 (* 0.0272727 = 0.117744 loss) | |
| I0506 00:21:18.209858 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 4.28913 (* 0.0272727 = 0.116976 loss) | |
| I0506 00:21:18.209872 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 4.37894 (* 0.0272727 = 0.119426 loss) | |
| I0506 00:21:18.209885 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 4.34273 (* 0.0272727 = 0.118438 loss) | |
| I0506 00:21:18.209899 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 4.28937 (* 0.0272727 = 0.116983 loss) | |
| I0506 00:21:18.209911 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:21:18.209923 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:21:18.209935 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 00:21:18.209946 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:21:18.209964 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 00:21:18.209975 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0506 00:21:18.209986 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0 | |
| I0506 00:21:18.209997 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0506 00:21:18.210010 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0 | |
| I0506 00:21:18.210029 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0 | |
| I0506 00:21:18.210041 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0 | |
| I0506 00:21:18.210052 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0 | |
| I0506 00:21:18.210062 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0 | |
| I0506 00:21:18.210074 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0 | |
| I0506 00:21:18.210085 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0 | |
| I0506 00:21:18.210109 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0 | |
| I0506 00:21:18.210122 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0 | |
| I0506 00:21:18.210134 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0 | |
| I0506 00:21:18.210145 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0 | |
| I0506 00:21:18.210156 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0 | |
| I0506 00:21:18.210167 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 0 | |
| I0506 00:21:18.210178 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 0 | |
| I0506 00:21:18.210189 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 0 | |
| I0506 00:21:18.210201 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0 | |
| I0506 00:21:18.210211 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0357143 | |
| I0506 00:21:18.210225 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.30228 (* 0.3 = 1.29068 loss) | |
| I0506 00:21:18.210239 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 4.29985 (* 0.3 = 1.28996 loss) | |
| I0506 00:21:18.210253 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.29906 (* 0.0272727 = 0.117247 loss) | |
| I0506 00:21:18.210268 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.29953 (* 0.0272727 = 0.11726 loss) | |
| I0506 00:21:18.210281 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 4.30392 (* 0.0272727 = 0.11738 loss) | |
| I0506 00:21:18.210294 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.31178 (* 0.0272727 = 0.117594 loss) | |
| I0506 00:21:18.210311 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 4.30406 (* 0.0272727 = 0.117383 loss) | |
| I0506 00:21:18.210326 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 4.30301 (* 0.0272727 = 0.117355 loss) | |
| I0506 00:21:18.210340 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 4.29 (* 0.0272727 = 0.117 loss) | |
| I0506 00:21:18.210353 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 4.30608 (* 0.0272727 = 0.117438 loss) | |
| I0506 00:21:18.210367 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 4.30373 (* 0.0272727 = 0.117375 loss) | |
| I0506 00:21:18.210381 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 4.3076 (* 0.0272727 = 0.11748 loss) | |
| I0506 00:21:18.210396 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 4.31231 (* 0.0272727 = 0.117608 loss) | |
| I0506 00:21:18.210408 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 4.3022 (* 0.0272727 = 0.117333 loss) | |
| I0506 00:21:18.210422 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 4.29722 (* 0.0272727 = 0.117197 loss) | |
| I0506 00:21:18.210435 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 4.30905 (* 0.0272727 = 0.11752 loss) | |
| I0506 00:21:18.210449 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 4.31153 (* 0.0272727 = 0.117587 loss) | |
| I0506 00:21:18.210464 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 4.31299 (* 0.0272727 = 0.117627 loss) | |
| I0506 00:21:18.210476 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 4.2969 (* 0.0272727 = 0.117188 loss) | |
| I0506 00:21:18.210490 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 4.29871 (* 0.0272727 = 0.117238 loss) | |
| I0506 00:21:18.210505 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 4.30277 (* 0.0272727 = 0.117348 loss) | |
| I0506 00:21:18.210518 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 4.31366 (* 0.0272727 = 0.117645 loss) | |
| I0506 00:21:18.210531 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 4.30117 (* 0.0272727 = 0.117305 loss) | |
| I0506 00:21:18.210544 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 4.30819 (* 0.0272727 = 0.117496 loss) | |
| I0506 00:21:18.210556 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0506 00:21:18.210575 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:21:18.210597 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:21:18.210610 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 00:21:18.210630 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 00:21:18.210642 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0506 00:21:18.210650 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0 | |
| I0506 00:21:18.210657 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0 | |
| I0506 00:21:18.210664 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0 | |
| I0506 00:21:18.210680 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.5 | |
| I0506 00:21:18.210700 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0 | |
| I0506 00:21:18.210711 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0 | |
| I0506 00:21:18.210737 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0 | |
| I0506 00:21:18.210752 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0 | |
| I0506 00:21:18.210764 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.5 | |
| I0506 00:21:18.210777 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.25 | |
| I0506 00:21:18.210796 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0 | |
| I0506 00:21:18.210808 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0 | |
| I0506 00:21:18.210819 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0 | |
| I0506 00:21:18.210829 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0 | |
| I0506 00:21:18.210840 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 0 | |
| I0506 00:21:18.210851 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 0.25 | |
| I0506 00:21:18.210867 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 0 | |
| I0506 00:21:18.210880 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0 | |
| I0506 00:21:18.210891 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0357143 | |
| I0506 00:21:18.210906 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.30345 (* 1 = 4.30345 loss) | |
| I0506 00:21:18.210919 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 4.30733 (* 1 = 4.30733 loss) | |
| I0506 00:21:18.210932 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 4.30537 (* 0.0909091 = 0.391397 loss) | |
| I0506 00:21:18.210947 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 4.30413 (* 0.0909091 = 0.391285 loss) | |
| I0506 00:21:18.210965 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 4.30106 (* 0.0909091 = 0.391006 loss) | |
| I0506 00:21:18.210979 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 4.30585 (* 0.0909091 = 0.391441 loss) | |
| I0506 00:21:18.210993 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 4.3096 (* 0.0909091 = 0.391781 loss) | |
| I0506 00:21:18.211005 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 4.30686 (* 0.0909091 = 0.391533 loss) | |
| I0506 00:21:18.211019 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 4.3029 (* 0.0909091 = 0.391173 loss) | |
| I0506 00:21:18.211037 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 4.30354 (* 0.0909091 = 0.391231 loss) | |
| I0506 00:21:18.211051 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 4.29065 (* 0.0909091 = 0.390059 loss) | |
| I0506 00:21:18.211064 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 4.30462 (* 0.0909091 = 0.391329 loss) | |
| I0506 00:21:18.211078 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 4.30367 (* 0.0909091 = 0.391242 loss) | |
| I0506 00:21:18.211091 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 4.30754 (* 0.0909091 = 0.391594 loss) | |
| I0506 00:21:18.211105 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 4.30465 (* 0.0909091 = 0.391332 loss) | |
| I0506 00:21:18.211118 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 4.28587 (* 0.0909091 = 0.389624 loss) | |
| I0506 00:21:18.211143 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 4.2886 (* 0.0909091 = 0.389873 loss) | |
| I0506 00:21:18.211158 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 4.30016 (* 0.0909091 = 0.390924 loss) | |
| I0506 00:21:18.211172 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 4.30987 (* 0.0909091 = 0.391807 loss) | |
| I0506 00:21:18.211186 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 4.29836 (* 0.0909091 = 0.39076 loss) | |
| I0506 00:21:18.211199 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 4.30481 (* 0.0909091 = 0.391347 loss) | |
| I0506 00:21:18.211213 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 4.30033 (* 0.0909091 = 0.39094 loss) | |
| I0506 00:21:18.211226 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 4.294 (* 0.0909091 = 0.390363 loss) | |
| I0506 00:21:18.211241 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 4.30833 (* 0.0909091 = 0.391667 loss) | |
| I0506 00:21:18.211252 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:21:18.211263 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:21:18.211274 15760 solver.cpp:245] Train net output #149: total_confidence = 1.07311e-41 | |
| I0506 00:21:18.211287 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.19166e-41 | |
| I0506 00:21:18.211308 15760 sgd_solver.cpp:106] Iteration 0, lr = 0.001 | |
| I0506 00:23:05.545322 15760 solver.cpp:229] Iteration 500, loss = 16.3829 | |
| I0506 00:23:05.545789 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:23:05.545810 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:23:05.545824 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:23:05.545835 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:23:05.545846 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:23:05.545858 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:23:05.545871 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 00:23:05.545886 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:23:05.545898 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 00:23:05.545910 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:23:05.545922 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:23:05.545934 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:23:05.545946 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 00:23:05.545958 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:23:05.545969 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:23:05.545981 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:23:05.545994 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:23:05.546005 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:23:05.546016 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:23:05.546028 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:23:05.546039 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:23:05.546051 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:23:05.546062 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:23:05.546074 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.693182 | |
| I0506 00:23:05.546087 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.111111 | |
| I0506 00:23:05.546103 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.56627 (* 0.3 = 1.36988 loss) | |
| I0506 00:23:05.546125 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.89838 (* 0.3 = 0.569514 loss) | |
| I0506 00:23:05.546140 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.87167 (* 0.0272727 = 0.132864 loss) | |
| I0506 00:23:05.546154 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.69127 (* 0.0272727 = 0.127944 loss) | |
| I0506 00:23:05.546169 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.39758 (* 0.0272727 = 0.119934 loss) | |
| I0506 00:23:05.546182 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 4.14132 (* 0.0272727 = 0.112945 loss) | |
| I0506 00:23:05.546195 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 4.16997 (* 0.0272727 = 0.113726 loss) | |
| I0506 00:23:05.546210 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.39378 (* 0.0272727 = 0.0925576 loss) | |
| I0506 00:23:05.546222 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 3.21706 (* 0.0272727 = 0.0877381 loss) | |
| I0506 00:23:05.546236 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 2.01055 (* 0.0272727 = 0.054833 loss) | |
| I0506 00:23:05.546250 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.874339 (* 0.0272727 = 0.0238456 loss) | |
| I0506 00:23:05.546264 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 1.03425 (* 0.0272727 = 0.0282069 loss) | |
| I0506 00:23:05.546277 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.963172 (* 0.0272727 = 0.0262683 loss) | |
| I0506 00:23:05.546291 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.887575 (* 0.0272727 = 0.0242066 loss) | |
| I0506 00:23:05.546306 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0540829 (* 0.0272727 = 0.00147499 loss) | |
| I0506 00:23:05.546339 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0535799 (* 0.0272727 = 0.00146127 loss) | |
| I0506 00:23:05.546355 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0688035 (* 0.0272727 = 0.00187646 loss) | |
| I0506 00:23:05.546368 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0584937 (* 0.0272727 = 0.00159528 loss) | |
| I0506 00:23:05.546381 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0604571 (* 0.0272727 = 0.00164883 loss) | |
| I0506 00:23:05.546396 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0340607 (* 0.0272727 = 0.000928929 loss) | |
| I0506 00:23:05.546409 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0290498 (* 0.0272727 = 0.000792267 loss) | |
| I0506 00:23:05.546423 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0359117 (* 0.0272727 = 0.000979409 loss) | |
| I0506 00:23:05.546437 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0465619 (* 0.0272727 = 0.00126987 loss) | |
| I0506 00:23:05.546452 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.029777 (* 0.0272727 = 0.0008121 loss) | |
| I0506 00:23:05.546464 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:23:05.546481 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:23:05.546504 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:23:05.546525 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:23:05.546538 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:23:05.546550 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 00:23:05.546561 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 00:23:05.546572 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:23:05.546584 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0506 00:23:05.546596 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:23:05.546608 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:23:05.546619 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:23:05.546633 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 00:23:05.546645 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:23:05.546658 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:23:05.546669 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:23:05.546680 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:23:05.546691 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:23:05.546702 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:23:05.546722 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:23:05.546735 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:23:05.546746 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:23:05.546757 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:23:05.546768 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0506 00:23:05.546780 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0185185 | |
| I0506 00:23:05.546794 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.46598 (* 0.3 = 1.33979 loss) | |
| I0506 00:23:05.546813 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.9682 (* 0.3 = 0.59046 loss) | |
| I0506 00:23:05.546828 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.35242 (* 0.0272727 = 0.118702 loss) | |
| I0506 00:23:05.546841 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.92942 (* 0.0272727 = 0.134439 loss) | |
| I0506 00:23:05.546867 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 5.21635 (* 0.0272727 = 0.142264 loss) | |
| I0506 00:23:05.546882 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.18059 (* 0.0272727 = 0.114016 loss) | |
| I0506 00:23:05.546896 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 4.10592 (* 0.0272727 = 0.11198 loss) | |
| I0506 00:23:05.546911 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 4.10223 (* 0.0272727 = 0.111879 loss) | |
| I0506 00:23:05.546923 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 3.87705 (* 0.0272727 = 0.105738 loss) | |
| I0506 00:23:05.546941 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 2.42071 (* 0.0272727 = 0.0660193 loss) | |
| I0506 00:23:05.546955 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.804887 (* 0.0272727 = 0.0219515 loss) | |
| I0506 00:23:05.546969 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.902877 (* 0.0272727 = 0.0246239 loss) | |
| I0506 00:23:05.546983 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.845956 (* 0.0272727 = 0.0230715 loss) | |
| I0506 00:23:05.546998 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 1.01545 (* 0.0272727 = 0.027694 loss) | |
| I0506 00:23:05.547011 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0807468 (* 0.0272727 = 0.00220219 loss) | |
| I0506 00:23:05.547025 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0782937 (* 0.0272727 = 0.00213528 loss) | |
| I0506 00:23:05.547039 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0928551 (* 0.0272727 = 0.00253241 loss) | |
| I0506 00:23:05.547054 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0373838 (* 0.0272727 = 0.00101956 loss) | |
| I0506 00:23:05.547066 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0381566 (* 0.0272727 = 0.00104063 loss) | |
| I0506 00:23:05.547080 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0533634 (* 0.0272727 = 0.00145537 loss) | |
| I0506 00:23:05.547094 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0300339 (* 0.0272727 = 0.000819106 loss) | |
| I0506 00:23:05.547107 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0274469 (* 0.0272727 = 0.000748553 loss) | |
| I0506 00:23:05.547122 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0792853 (* 0.0272727 = 0.00216233 loss) | |
| I0506 00:23:05.547135 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0315272 (* 0.0272727 = 0.000859833 loss) | |
| I0506 00:23:05.547147 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0506 00:23:05.547158 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 00:23:05.547170 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:23:05.547183 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:23:05.547194 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:23:05.547205 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:23:05.547217 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 00:23:05.547230 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:23:05.547241 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0506 00:23:05.547252 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:23:05.547265 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:23:05.547276 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:23:05.547287 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 00:23:05.547299 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:23:05.547310 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:23:05.547322 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:23:05.547333 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:23:05.547354 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:23:05.547368 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:23:05.547379 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:23:05.547389 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:23:05.547400 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:23:05.547412 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:23:05.547423 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.693182 | |
| I0506 00:23:05.547435 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.037037 | |
| I0506 00:23:05.547449 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.38273 (* 1 = 4.38273 loss) | |
| I0506 00:23:05.547463 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.53588 (* 1 = 1.53588 loss) | |
| I0506 00:23:05.547477 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 4.0189 (* 0.0909091 = 0.365355 loss) | |
| I0506 00:23:05.547490 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 4.11131 (* 0.0909091 = 0.373756 loss) | |
| I0506 00:23:05.547504 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 4.03544 (* 0.0909091 = 0.366858 loss) | |
| I0506 00:23:05.547518 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.84076 (* 0.0909091 = 0.34916 loss) | |
| I0506 00:23:05.547531 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 4.0625 (* 0.0909091 = 0.369318 loss) | |
| I0506 00:23:05.547544 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.62582 (* 0.0909091 = 0.32962 loss) | |
| I0506 00:23:05.547559 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.90576 (* 0.0909091 = 0.26416 loss) | |
| I0506 00:23:05.547572 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 2.45031 (* 0.0909091 = 0.222756 loss) | |
| I0506 00:23:05.547585 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.859863 (* 0.0909091 = 0.0781693 loss) | |
| I0506 00:23:05.547598 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.808726 (* 0.0909091 = 0.0735206 loss) | |
| I0506 00:23:05.547612 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.74866 (* 0.0909091 = 0.06806 loss) | |
| I0506 00:23:05.547626 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.707002 (* 0.0909091 = 0.0642729 loss) | |
| I0506 00:23:05.547639 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0328685 (* 0.0909091 = 0.00298805 loss) | |
| I0506 00:23:05.547654 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0287749 (* 0.0909091 = 0.0026159 loss) | |
| I0506 00:23:05.547668 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0201702 (* 0.0909091 = 0.00183365 loss) | |
| I0506 00:23:05.547683 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.011708 (* 0.0909091 = 0.00106436 loss) | |
| I0506 00:23:05.547696 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00883825 (* 0.0909091 = 0.000803478 loss) | |
| I0506 00:23:05.547710 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00536457 (* 0.0909091 = 0.000487688 loss) | |
| I0506 00:23:05.547724 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00964439 (* 0.0909091 = 0.000876762 loss) | |
| I0506 00:23:05.547737 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00593543 (* 0.0909091 = 0.000539584 loss) | |
| I0506 00:23:05.547751 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0042607 (* 0.0909091 = 0.000387336 loss) | |
| I0506 00:23:05.547765 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00615515 (* 0.0909091 = 0.000559559 loss) | |
| I0506 00:23:05.547776 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:23:05.547788 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:23:05.547799 15760 solver.cpp:245] Train net output #149: total_confidence = 1.75176e-08 | |
| I0506 00:23:05.547821 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.24725e-05 | |
| I0506 00:23:05.547835 15760 sgd_solver.cpp:106] Iteration 500, lr = 0.001 | |
| I0506 00:24:52.640451 15760 solver.cpp:229] Iteration 1000, loss = 13.3096 | |
| I0506 00:24:52.640609 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:24:52.640630 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:24:52.640642 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:24:52.640655 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 00:24:52.640666 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:24:52.640678 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:24:52.640689 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 00:24:52.640702 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:24:52.640713 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:24:52.640725 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:24:52.640738 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:24:52.640749 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:24:52.640760 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 00:24:52.640772 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 00:24:52.640784 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 00:24:52.640796 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:24:52.640808 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:24:52.640820 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:24:52.640832 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:24:52.640843 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:24:52.640856 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:24:52.640866 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:24:52.640882 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:24:52.640893 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.681818 | |
| I0506 00:24:52.640905 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0714286 | |
| I0506 00:24:52.640923 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.09987 (* 0.3 = 1.22996 loss) | |
| I0506 00:24:52.640945 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.77853 (* 0.3 = 0.533559 loss) | |
| I0506 00:24:52.640960 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.74897 (* 0.0272727 = 0.102245 loss) | |
| I0506 00:24:52.640974 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.52189 (* 0.0272727 = 0.123324 loss) | |
| I0506 00:24:52.640988 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.21062 (* 0.0272727 = 0.114835 loss) | |
| I0506 00:24:52.641002 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.9198 (* 0.0272727 = 0.106904 loss) | |
| I0506 00:24:52.641016 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.52713 (* 0.0272727 = 0.0961944 loss) | |
| I0506 00:24:52.641031 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.72529 (* 0.0272727 = 0.101599 loss) | |
| I0506 00:24:52.641043 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 3.3599 (* 0.0272727 = 0.0916337 loss) | |
| I0506 00:24:52.641057 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.18466 (* 0.0272727 = 0.032309 loss) | |
| I0506 00:24:52.641070 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.35041 (* 0.0272727 = 0.0368293 loss) | |
| I0506 00:24:52.641084 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 1.24964 (* 0.0272727 = 0.0340812 loss) | |
| I0506 00:24:52.641098 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 1.22607 (* 0.0272727 = 0.0334383 loss) | |
| I0506 00:24:52.641111 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 1.37914 (* 0.0272727 = 0.037613 loss) | |
| I0506 00:24:52.641144 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 1.22804 (* 0.0272727 = 0.033492 loss) | |
| I0506 00:24:52.641180 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 1.44944 (* 0.0272727 = 0.0395303 loss) | |
| I0506 00:24:52.641196 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0695762 (* 0.0272727 = 0.00189753 loss) | |
| I0506 00:24:52.641211 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0420125 (* 0.0272727 = 0.0011458 loss) | |
| I0506 00:24:52.641225 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0312006 (* 0.0272727 = 0.000850926 loss) | |
| I0506 00:24:52.641238 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0325085 (* 0.0272727 = 0.000886596 loss) | |
| I0506 00:24:52.641252 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0374326 (* 0.0272727 = 0.00102089 loss) | |
| I0506 00:24:52.641266 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0392999 (* 0.0272727 = 0.00107182 loss) | |
| I0506 00:24:52.641280 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0385019 (* 0.0272727 = 0.00105005 loss) | |
| I0506 00:24:52.641294 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0372033 (* 0.0272727 = 0.00101464 loss) | |
| I0506 00:24:52.641306 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:24:52.641317 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:24:52.641330 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:24:52.641340 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:24:52.641351 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 00:24:52.641362 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0506 00:24:52.641373 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 00:24:52.641386 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:24:52.641396 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:24:52.641408 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:24:52.641419 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:24:52.641432 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:24:52.641443 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 00:24:52.641454 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 00:24:52.641466 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 00:24:52.641479 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:24:52.641485 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:24:52.641494 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:24:52.641500 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:24:52.641512 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:24:52.641525 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:24:52.641535 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:24:52.641547 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:24:52.641564 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.681818 | |
| I0506 00:24:52.641577 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.160714 | |
| I0506 00:24:52.641592 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.00529 (* 0.3 = 1.20159 loss) | |
| I0506 00:24:52.641607 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.76381 (* 0.3 = 0.529142 loss) | |
| I0506 00:24:52.641623 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.29422 (* 0.0272727 = 0.117115 loss) | |
| I0506 00:24:52.641638 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.04607 (* 0.0272727 = 0.110347 loss) | |
| I0506 00:24:52.641652 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 4.31008 (* 0.0272727 = 0.117548 loss) | |
| I0506 00:24:52.641677 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.98193 (* 0.0272727 = 0.108598 loss) | |
| I0506 00:24:52.641692 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 4.43462 (* 0.0272727 = 0.120944 loss) | |
| I0506 00:24:52.641711 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.84968 (* 0.0272727 = 0.104991 loss) | |
| I0506 00:24:52.641726 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.7958 (* 0.0272727 = 0.0762491 loss) | |
| I0506 00:24:52.641739 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.96776 (* 0.0272727 = 0.0263934 loss) | |
| I0506 00:24:52.641752 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.01108 (* 0.0272727 = 0.0275749 loss) | |
| I0506 00:24:52.641767 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 1.01574 (* 0.0272727 = 0.0277021 loss) | |
| I0506 00:24:52.641779 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.999028 (* 0.0272727 = 0.0272462 loss) | |
| I0506 00:24:52.641793 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 1.18247 (* 0.0272727 = 0.0322492 loss) | |
| I0506 00:24:52.641806 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.787272 (* 0.0272727 = 0.0214711 loss) | |
| I0506 00:24:52.641820 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 1.06647 (* 0.0272727 = 0.0290856 loss) | |
| I0506 00:24:52.641834 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0389922 (* 0.0272727 = 0.00106342 loss) | |
| I0506 00:24:52.641847 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0151379 (* 0.0272727 = 0.000412852 loss) | |
| I0506 00:24:52.641861 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0169646 (* 0.0272727 = 0.00046267 loss) | |
| I0506 00:24:52.641875 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0195236 (* 0.0272727 = 0.000532462 loss) | |
| I0506 00:24:52.641888 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0115629 (* 0.0272727 = 0.000315351 loss) | |
| I0506 00:24:52.641902 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0199931 (* 0.0272727 = 0.000545266 loss) | |
| I0506 00:24:52.641916 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0198533 (* 0.0272727 = 0.000541453 loss) | |
| I0506 00:24:52.641932 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0276728 (* 0.0272727 = 0.000754714 loss) | |
| I0506 00:24:52.641945 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0535714 | |
| I0506 00:24:52.641957 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:24:52.641968 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:24:52.641979 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:24:52.641990 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 00:24:52.642001 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:24:52.642014 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 00:24:52.642024 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:24:52.642036 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:24:52.642047 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:24:52.642060 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:24:52.642071 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:24:52.642082 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 00:24:52.642093 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 00:24:52.642105 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 00:24:52.642117 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:24:52.642128 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:24:52.642149 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:24:52.642163 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:24:52.642174 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:24:52.642184 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:24:52.642196 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:24:52.642207 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:24:52.642218 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0506 00:24:52.642230 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.125 | |
| I0506 00:24:52.642243 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.8648 (* 1 = 3.8648 loss) | |
| I0506 00:24:52.642256 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.47553 (* 1 = 1.47553 loss) | |
| I0506 00:24:52.642271 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.60432 (* 0.0909091 = 0.327666 loss) | |
| I0506 00:24:52.642284 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.73804 (* 0.0909091 = 0.339822 loss) | |
| I0506 00:24:52.642297 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 4.18159 (* 0.0909091 = 0.380145 loss) | |
| I0506 00:24:52.642312 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.61211 (* 0.0909091 = 0.328374 loss) | |
| I0506 00:24:52.642324 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.73111 (* 0.0909091 = 0.339192 loss) | |
| I0506 00:24:52.642338 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.55194 (* 0.0909091 = 0.322904 loss) | |
| I0506 00:24:52.642351 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.42822 (* 0.0909091 = 0.220748 loss) | |
| I0506 00:24:52.642364 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.892041 (* 0.0909091 = 0.0810946 loss) | |
| I0506 00:24:52.642379 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.780486 (* 0.0909091 = 0.0709532 loss) | |
| I0506 00:24:52.642391 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.878631 (* 0.0909091 = 0.0798756 loss) | |
| I0506 00:24:52.642405 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 1.03265 (* 0.0909091 = 0.0938774 loss) | |
| I0506 00:24:52.642418 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 1.11645 (* 0.0909091 = 0.101496 loss) | |
| I0506 00:24:52.642436 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 1.053 (* 0.0909091 = 0.0957275 loss) | |
| I0506 00:24:52.642462 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 1.17701 (* 0.0909091 = 0.107001 loss) | |
| I0506 00:24:52.642482 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00800252 (* 0.0909091 = 0.000727502 loss) | |
| I0506 00:24:52.642495 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00469739 (* 0.0909091 = 0.000427036 loss) | |
| I0506 00:24:52.642510 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00443574 (* 0.0909091 = 0.000403249 loss) | |
| I0506 00:24:52.642524 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00316208 (* 0.0909091 = 0.000287462 loss) | |
| I0506 00:24:52.642537 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00269565 (* 0.0909091 = 0.000245059 loss) | |
| I0506 00:24:52.642551 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00220258 (* 0.0909091 = 0.000200235 loss) | |
| I0506 00:24:52.642565 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00232497 (* 0.0909091 = 0.000211361 loss) | |
| I0506 00:24:52.642578 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00257904 (* 0.0909091 = 0.000234458 loss) | |
| I0506 00:24:52.642590 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:24:52.642601 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:24:52.642612 15760 solver.cpp:245] Train net output #149: total_confidence = 4.59729e-09 | |
| I0506 00:24:52.642635 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.06196e-06 | |
| I0506 00:24:52.642650 15760 sgd_solver.cpp:106] Iteration 1000, lr = 0.001 | |
| I0506 00:26:39.756435 15760 solver.cpp:229] Iteration 1500, loss = 13.0559 | |
| I0506 00:26:39.756608 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.025641 | |
| I0506 00:26:39.756628 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:26:39.756641 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 00:26:39.756654 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:26:39.756665 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 00:26:39.756677 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.625 | |
| I0506 00:26:39.756690 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0506 00:26:39.756701 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 00:26:39.756713 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 00:26:39.756724 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:26:39.756736 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:26:39.756747 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:26:39.756759 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:26:39.756770 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:26:39.756783 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:26:39.756793 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:26:39.756805 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:26:39.756817 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:26:39.756829 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:26:39.756840 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:26:39.756852 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:26:39.756863 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:26:39.756878 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:26:39.756891 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.784091 | |
| I0506 00:26:39.756902 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.128205 | |
| I0506 00:26:39.756919 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.98222 (* 0.3 = 1.19466 loss) | |
| I0506 00:26:39.756933 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.2733 (* 0.3 = 0.381991 loss) | |
| I0506 00:26:39.756947 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.17234 (* 0.0272727 = 0.113791 loss) | |
| I0506 00:26:39.756963 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.98602 (* 0.0272727 = 0.10871 loss) | |
| I0506 00:26:39.756975 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.07679 (* 0.0272727 = 0.111185 loss) | |
| I0506 00:26:39.756989 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.99092 (* 0.0272727 = 0.108843 loss) | |
| I0506 00:26:39.757004 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.49714 (* 0.0272727 = 0.0681039 loss) | |
| I0506 00:26:39.757016 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.5788 (* 0.0272727 = 0.0430582 loss) | |
| I0506 00:26:39.757030 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.61359 (* 0.0272727 = 0.044007 loss) | |
| I0506 00:26:39.757043 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.35327 (* 0.0272727 = 0.0369075 loss) | |
| I0506 00:26:39.757058 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.139368 (* 0.0272727 = 0.00380094 loss) | |
| I0506 00:26:39.757072 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.127891 (* 0.0272727 = 0.00348793 loss) | |
| I0506 00:26:39.757086 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.146241 (* 0.0272727 = 0.0039884 loss) | |
| I0506 00:26:39.757100 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.110602 (* 0.0272727 = 0.00301642 loss) | |
| I0506 00:26:39.757114 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.129875 (* 0.0272727 = 0.00354205 loss) | |
| I0506 00:26:39.757164 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0834726 (* 0.0272727 = 0.00227652 loss) | |
| I0506 00:26:39.757180 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0744574 (* 0.0272727 = 0.00203066 loss) | |
| I0506 00:26:39.757194 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0653196 (* 0.0272727 = 0.00178144 loss) | |
| I0506 00:26:39.757208 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0443507 (* 0.0272727 = 0.00120956 loss) | |
| I0506 00:26:39.757222 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0473176 (* 0.0272727 = 0.00129048 loss) | |
| I0506 00:26:39.757236 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0426922 (* 0.0272727 = 0.00116433 loss) | |
| I0506 00:26:39.757249 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0366324 (* 0.0272727 = 0.000999064 loss) | |
| I0506 00:26:39.757263 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.049832 (* 0.0272727 = 0.00135906 loss) | |
| I0506 00:26:39.757277 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.051196 (* 0.0272727 = 0.00139626 loss) | |
| I0506 00:26:39.757289 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:26:39.757302 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:26:39.757313 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:26:39.757324 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:26:39.757335 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:26:39.757347 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.625 | |
| I0506 00:26:39.757359 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0506 00:26:39.757371 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 00:26:39.757382 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 00:26:39.757393 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:26:39.757405 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:26:39.757416 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:26:39.757427 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:26:39.757438 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:26:39.757449 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:26:39.757462 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:26:39.757472 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:26:39.757484 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:26:39.757495 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:26:39.757508 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:26:39.757519 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:26:39.757530 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:26:39.757541 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:26:39.757552 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0506 00:26:39.757565 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0769231 | |
| I0506 00:26:39.757585 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.04705 (* 0.3 = 1.21411 loss) | |
| I0506 00:26:39.757614 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.36421 (* 0.3 = 0.409264 loss) | |
| I0506 00:26:39.757637 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.02083 (* 0.0272727 = 0.109659 loss) | |
| I0506 00:26:39.757652 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.00307 (* 0.0272727 = 0.109175 loss) | |
| I0506 00:26:39.757665 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.92114 (* 0.0272727 = 0.10694 loss) | |
| I0506 00:26:39.757693 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.95336 (* 0.0272727 = 0.107819 loss) | |
| I0506 00:26:39.757707 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.81371 (* 0.0272727 = 0.0767374 loss) | |
| I0506 00:26:39.757721 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.26678 (* 0.0272727 = 0.0618214 loss) | |
| I0506 00:26:39.757735 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.55512 (* 0.0272727 = 0.0424123 loss) | |
| I0506 00:26:39.757748 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.59181 (* 0.0272727 = 0.043413 loss) | |
| I0506 00:26:39.757762 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.231361 (* 0.0272727 = 0.00630983 loss) | |
| I0506 00:26:39.757776 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.135195 (* 0.0272727 = 0.00368713 loss) | |
| I0506 00:26:39.757789 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.121192 (* 0.0272727 = 0.00330524 loss) | |
| I0506 00:26:39.757803 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.104785 (* 0.0272727 = 0.00285777 loss) | |
| I0506 00:26:39.757817 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.158118 (* 0.0272727 = 0.00431231 loss) | |
| I0506 00:26:39.757832 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.100016 (* 0.0272727 = 0.0027277 loss) | |
| I0506 00:26:39.757845 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0704425 (* 0.0272727 = 0.00192116 loss) | |
| I0506 00:26:39.757859 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0454905 (* 0.0272727 = 0.00124065 loss) | |
| I0506 00:26:39.757874 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0427882 (* 0.0272727 = 0.00116695 loss) | |
| I0506 00:26:39.757887 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0650557 (* 0.0272727 = 0.00177425 loss) | |
| I0506 00:26:39.757901 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.041326 (* 0.0272727 = 0.00112707 loss) | |
| I0506 00:26:39.757915 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0323184 (* 0.0272727 = 0.00088141 loss) | |
| I0506 00:26:39.757931 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0558574 (* 0.0272727 = 0.00152338 loss) | |
| I0506 00:26:39.757946 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0533589 (* 0.0272727 = 0.00145524 loss) | |
| I0506 00:26:39.757958 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0506 00:26:39.757971 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:26:39.757982 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:26:39.757992 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:26:39.758004 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 00:26:39.758015 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0506 00:26:39.758028 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.75 | |
| I0506 00:26:39.758038 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 00:26:39.758049 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 00:26:39.758061 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:26:39.758074 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:26:39.758085 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:26:39.758095 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:26:39.758106 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:26:39.758117 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:26:39.758129 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:26:39.758141 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:26:39.758152 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:26:39.758173 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:26:39.758186 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:26:39.758198 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:26:39.758208 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:26:39.758220 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:26:39.758231 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0506 00:26:39.758242 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0512821 | |
| I0506 00:26:39.758256 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.71762 (* 1 = 3.71762 loss) | |
| I0506 00:26:39.758270 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.20591 (* 1 = 1.20591 loss) | |
| I0506 00:26:39.758283 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.84848 (* 0.0909091 = 0.349862 loss) | |
| I0506 00:26:39.758297 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.74599 (* 0.0909091 = 0.340544 loss) | |
| I0506 00:26:39.758311 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.96269 (* 0.0909091 = 0.360244 loss) | |
| I0506 00:26:39.758324 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.44116 (* 0.0909091 = 0.312833 loss) | |
| I0506 00:26:39.758338 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.14927 (* 0.0909091 = 0.195388 loss) | |
| I0506 00:26:39.758352 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.46511 (* 0.0909091 = 0.133192 loss) | |
| I0506 00:26:39.758364 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.20192 (* 0.0909091 = 0.109266 loss) | |
| I0506 00:26:39.758378 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.34568 (* 0.0909091 = 0.122334 loss) | |
| I0506 00:26:39.758393 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0982478 (* 0.0909091 = 0.00893162 loss) | |
| I0506 00:26:39.758406 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0791269 (* 0.0909091 = 0.00719336 loss) | |
| I0506 00:26:39.758419 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0715274 (* 0.0909091 = 0.00650249 loss) | |
| I0506 00:26:39.758433 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0571588 (* 0.0909091 = 0.00519625 loss) | |
| I0506 00:26:39.758447 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0405384 (* 0.0909091 = 0.00368531 loss) | |
| I0506 00:26:39.758461 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0327443 (* 0.0909091 = 0.00297675 loss) | |
| I0506 00:26:39.758474 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.027085 (* 0.0909091 = 0.00246228 loss) | |
| I0506 00:26:39.758488 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0170948 (* 0.0909091 = 0.00155407 loss) | |
| I0506 00:26:39.758502 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00911789 (* 0.0909091 = 0.000828899 loss) | |
| I0506 00:26:39.758515 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0114029 (* 0.0909091 = 0.00103663 loss) | |
| I0506 00:26:39.758529 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00997418 (* 0.0909091 = 0.000906744 loss) | |
| I0506 00:26:39.758543 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00685661 (* 0.0909091 = 0.000623329 loss) | |
| I0506 00:26:39.758556 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00704779 (* 0.0909091 = 0.000640708 loss) | |
| I0506 00:26:39.758570 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00887051 (* 0.0909091 = 0.00080641 loss) | |
| I0506 00:26:39.758582 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:26:39.758594 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:26:39.758605 15760 solver.cpp:245] Train net output #149: total_confidence = 2.01951e-08 | |
| I0506 00:26:39.758615 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.88337e-05 | |
| I0506 00:26:39.758638 15760 sgd_solver.cpp:106] Iteration 1500, lr = 0.001 | |
| I0506 00:28:26.846534 15760 solver.cpp:229] Iteration 2000, loss = 12.9355 | |
| I0506 00:28:26.846665 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:28:26.846686 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:28:26.846698 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:28:26.846710 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:28:26.846721 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 00:28:26.846735 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:28:26.846745 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 00:28:26.846757 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:28:26.846770 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 00:28:26.846781 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:28:26.846793 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:28:26.846804 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:28:26.846817 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:28:26.846827 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:28:26.846839 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:28:26.846850 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:28:26.846869 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:28:26.846885 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:28:26.846897 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:28:26.846909 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:28:26.846920 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:28:26.846931 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:28:26.846942 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:28:26.846953 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0506 00:28:26.846966 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0588235 | |
| I0506 00:28:26.846982 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.58497 (* 0.3 = 1.37549 loss) | |
| I0506 00:28:26.846997 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.66613 (* 0.3 = 0.49984 loss) | |
| I0506 00:28:26.847010 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.25839 (* 0.0272727 = 0.116138 loss) | |
| I0506 00:28:26.847023 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.05794 (* 0.0272727 = 0.110671 loss) | |
| I0506 00:28:26.847038 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.22501 (* 0.0272727 = 0.115228 loss) | |
| I0506 00:28:26.847050 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.99626 (* 0.0272727 = 0.108989 loss) | |
| I0506 00:28:26.847064 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 4.50855 (* 0.0272727 = 0.12296 loss) | |
| I0506 00:28:26.847077 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.77043 (* 0.0272727 = 0.10283 loss) | |
| I0506 00:28:26.847090 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.85657 (* 0.0272727 = 0.0779064 loss) | |
| I0506 00:28:26.847105 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.77322 (* 0.0272727 = 0.0483606 loss) | |
| I0506 00:28:26.847118 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.06149 (* 0.0272727 = 0.0289497 loss) | |
| I0506 00:28:26.847132 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0835307 (* 0.0272727 = 0.00227811 loss) | |
| I0506 00:28:26.847147 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0625388 (* 0.0272727 = 0.0017056 loss) | |
| I0506 00:28:26.847162 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0419291 (* 0.0272727 = 0.00114352 loss) | |
| I0506 00:28:26.847180 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0543077 (* 0.0272727 = 0.00148112 loss) | |
| I0506 00:28:26.847213 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0316339 (* 0.0272727 = 0.000862741 loss) | |
| I0506 00:28:26.847229 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0301841 (* 0.0272727 = 0.000823203 loss) | |
| I0506 00:28:26.847242 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0259126 (* 0.0272727 = 0.000706706 loss) | |
| I0506 00:28:26.847256 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0161531 (* 0.0272727 = 0.000440539 loss) | |
| I0506 00:28:26.847270 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0375606 (* 0.0272727 = 0.00102438 loss) | |
| I0506 00:28:26.847283 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0132868 (* 0.0272727 = 0.000362368 loss) | |
| I0506 00:28:26.847297 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0164362 (* 0.0272727 = 0.000448261 loss) | |
| I0506 00:28:26.847311 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.016404 (* 0.0272727 = 0.00044738 loss) | |
| I0506 00:28:26.847324 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0144294 (* 0.0272727 = 0.000393529 loss) | |
| I0506 00:28:26.847337 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:28:26.847348 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:28:26.847359 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:28:26.847370 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:28:26.847381 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:28:26.847393 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 00:28:26.847404 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 00:28:26.847415 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:28:26.847426 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 00:28:26.847437 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:28:26.847450 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:28:26.847460 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:28:26.847471 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:28:26.847482 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:28:26.847492 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:28:26.847503 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:28:26.847514 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:28:26.847525 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:28:26.847537 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:28:26.847548 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:28:26.847558 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:28:26.847569 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:28:26.847580 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:28:26.847591 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0506 00:28:26.847602 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0196078 | |
| I0506 00:28:26.847616 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.75247 (* 0.3 = 1.42574 loss) | |
| I0506 00:28:26.847630 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.5897 (* 0.3 = 0.476909 loss) | |
| I0506 00:28:26.847642 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.30272 (* 0.0272727 = 0.117347 loss) | |
| I0506 00:28:26.847656 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.10631 (* 0.0272727 = 0.11199 loss) | |
| I0506 00:28:26.847669 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 4.15888 (* 0.0272727 = 0.113424 loss) | |
| I0506 00:28:26.847698 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.01243 (* 0.0272727 = 0.10943 loss) | |
| I0506 00:28:26.847714 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 4.58324 (* 0.0272727 = 0.124997 loss) | |
| I0506 00:28:26.847728 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.64666 (* 0.0272727 = 0.0994544 loss) | |
| I0506 00:28:26.847741 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 3.07906 (* 0.0272727 = 0.0839743 loss) | |
| I0506 00:28:26.847754 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.50057 (* 0.0272727 = 0.0409247 loss) | |
| I0506 00:28:26.847769 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.05007 (* 0.0272727 = 0.0286382 loss) | |
| I0506 00:28:26.847781 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0978743 (* 0.0272727 = 0.0026693 loss) | |
| I0506 00:28:26.847795 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0688154 (* 0.0272727 = 0.00187678 loss) | |
| I0506 00:28:26.847808 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0788556 (* 0.0272727 = 0.00215061 loss) | |
| I0506 00:28:26.847822 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0471807 (* 0.0272727 = 0.00128675 loss) | |
| I0506 00:28:26.847836 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0561733 (* 0.0272727 = 0.001532 loss) | |
| I0506 00:28:26.847851 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0381294 (* 0.0272727 = 0.00103989 loss) | |
| I0506 00:28:26.847863 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0364693 (* 0.0272727 = 0.000994617 loss) | |
| I0506 00:28:26.847877 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0223451 (* 0.0272727 = 0.000609412 loss) | |
| I0506 00:28:26.847892 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0204029 (* 0.0272727 = 0.000556443 loss) | |
| I0506 00:28:26.847904 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0276189 (* 0.0272727 = 0.000753243 loss) | |
| I0506 00:28:26.847918 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0186221 (* 0.0272727 = 0.000507875 loss) | |
| I0506 00:28:26.847935 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0113354 (* 0.0272727 = 0.000309147 loss) | |
| I0506 00:28:26.847949 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0148253 (* 0.0272727 = 0.000404328 loss) | |
| I0506 00:28:26.847961 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0196078 | |
| I0506 00:28:26.847973 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:28:26.847985 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 00:28:26.847996 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:28:26.848007 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:28:26.848018 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:28:26.848029 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 00:28:26.848040 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:28:26.848052 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 00:28:26.848063 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:28:26.848074 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:28:26.848085 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:28:26.848096 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:28:26.848107 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:28:26.848119 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:28:26.848130 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:28:26.848137 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:28:26.848145 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:28:26.848166 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:28:26.848179 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:28:26.848191 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:28:26.848201 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:28:26.848212 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:28:26.848223 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.715909 | |
| I0506 00:28:26.848234 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0784314 | |
| I0506 00:28:26.848248 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.56817 (* 1 = 4.56817 loss) | |
| I0506 00:28:26.848260 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.41079 (* 1 = 1.41079 loss) | |
| I0506 00:28:26.848274 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 4.14843 (* 0.0909091 = 0.37713 loss) | |
| I0506 00:28:26.848287 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.89591 (* 0.0909091 = 0.354174 loss) | |
| I0506 00:28:26.848300 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 4.32905 (* 0.0909091 = 0.39355 loss) | |
| I0506 00:28:26.848314 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 4.07549 (* 0.0909091 = 0.370499 loss) | |
| I0506 00:28:26.848326 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 4.20437 (* 0.0909091 = 0.382216 loss) | |
| I0506 00:28:26.848340 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.7577 (* 0.0909091 = 0.341609 loss) | |
| I0506 00:28:26.848352 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.89969 (* 0.0909091 = 0.263608 loss) | |
| I0506 00:28:26.848366 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.57 (* 0.0909091 = 0.142728 loss) | |
| I0506 00:28:26.848379 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.831506 (* 0.0909091 = 0.0755915 loss) | |
| I0506 00:28:26.848392 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.13764 (* 0.0909091 = 0.0125127 loss) | |
| I0506 00:28:26.848405 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.096592 (* 0.0909091 = 0.0087811 loss) | |
| I0506 00:28:26.848418 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.076536 (* 0.0909091 = 0.00695782 loss) | |
| I0506 00:28:26.848433 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0663221 (* 0.0909091 = 0.00602928 loss) | |
| I0506 00:28:26.848445 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0578812 (* 0.0909091 = 0.00526193 loss) | |
| I0506 00:28:26.848459 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0412424 (* 0.0909091 = 0.00374931 loss) | |
| I0506 00:28:26.848471 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0260153 (* 0.0909091 = 0.00236502 loss) | |
| I0506 00:28:26.848485 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0133953 (* 0.0909091 = 0.00121775 loss) | |
| I0506 00:28:26.848498 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0105016 (* 0.0909091 = 0.000954689 loss) | |
| I0506 00:28:26.848511 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0116866 (* 0.0909091 = 0.00106242 loss) | |
| I0506 00:28:26.848526 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00820997 (* 0.0909091 = 0.000746361 loss) | |
| I0506 00:28:26.848538 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00875113 (* 0.0909091 = 0.000795557 loss) | |
| I0506 00:28:26.848551 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00843555 (* 0.0909091 = 0.000766868 loss) | |
| I0506 00:28:26.848563 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:28:26.848573 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:28:26.848584 15760 solver.cpp:245] Train net output #149: total_confidence = 4.16044e-07 | |
| I0506 00:28:26.848597 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 4.35956e-06 | |
| I0506 00:28:26.848619 15760 sgd_solver.cpp:106] Iteration 2000, lr = 0.001 | |
| I0506 00:30:13.732159 15760 solver.cpp:229] Iteration 2500, loss = 12.7455 | |
| I0506 00:30:13.732319 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.025 | |
| I0506 00:30:13.732339 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:30:13.732353 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 00:30:13.732367 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:30:13.732378 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 00:30:13.732389 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0506 00:30:13.732401 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 00:30:13.732414 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 00:30:13.732425 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:30:13.732437 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:30:13.732450 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:30:13.732461 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:30:13.732472 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:30:13.732484 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:30:13.732496 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:30:13.732507 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:30:13.732518 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:30:13.732530 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:30:13.732542 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:30:13.732554 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:30:13.732565 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:30:13.732578 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:30:13.732589 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:30:13.732600 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0506 00:30:13.732612 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.025 | |
| I0506 00:30:13.732628 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.97107 (* 0.3 = 1.19132 loss) | |
| I0506 00:30:13.732642 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.42762 (* 0.3 = 0.428285 loss) | |
| I0506 00:30:13.732656 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.16652 (* 0.0272727 = 0.113632 loss) | |
| I0506 00:30:13.732671 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.07344 (* 0.0272727 = 0.111094 loss) | |
| I0506 00:30:13.732684 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.84805 (* 0.0272727 = 0.104947 loss) | |
| I0506 00:30:13.732698 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.29731 (* 0.0272727 = 0.0899265 loss) | |
| I0506 00:30:13.732712 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.91242 (* 0.0272727 = 0.0794295 loss) | |
| I0506 00:30:13.732727 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.17943 (* 0.0272727 = 0.0594389 loss) | |
| I0506 00:30:13.732739 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.5637 (* 0.0272727 = 0.0426464 loss) | |
| I0506 00:30:13.732753 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.967264 (* 0.0272727 = 0.0263799 loss) | |
| I0506 00:30:13.732767 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.86543 (* 0.0272727 = 0.0236026 loss) | |
| I0506 00:30:13.732781 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.111245 (* 0.0272727 = 0.00303396 loss) | |
| I0506 00:30:13.732795 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.143507 (* 0.0272727 = 0.00391384 loss) | |
| I0506 00:30:13.732810 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.10044 (* 0.0272727 = 0.00273927 loss) | |
| I0506 00:30:13.732823 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0798597 (* 0.0272727 = 0.00217799 loss) | |
| I0506 00:30:13.732857 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0651934 (* 0.0272727 = 0.001778 loss) | |
| I0506 00:30:13.732872 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.056648 (* 0.0272727 = 0.00154495 loss) | |
| I0506 00:30:13.732890 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0550949 (* 0.0272727 = 0.00150259 loss) | |
| I0506 00:30:13.732905 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0502941 (* 0.0272727 = 0.00137166 loss) | |
| I0506 00:30:13.732919 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.040641 (* 0.0272727 = 0.00110839 loss) | |
| I0506 00:30:13.732933 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0217395 (* 0.0272727 = 0.000592896 loss) | |
| I0506 00:30:13.732947 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0229452 (* 0.0272727 = 0.000625779 loss) | |
| I0506 00:30:13.732961 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0396954 (* 0.0272727 = 0.0010826 loss) | |
| I0506 00:30:13.732976 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0391358 (* 0.0272727 = 0.00106734 loss) | |
| I0506 00:30:13.732988 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.025 | |
| I0506 00:30:13.733000 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:30:13.733012 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:30:13.733024 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 00:30:13.733036 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0506 00:30:13.733047 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 00:30:13.733059 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 00:30:13.733070 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 00:30:13.733083 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:30:13.733093 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:30:13.733104 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:30:13.733116 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:30:13.733149 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:30:13.733160 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:30:13.733172 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:30:13.733185 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:30:13.733192 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:30:13.733199 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:30:13.733212 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:30:13.733224 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:30:13.733237 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:30:13.733247 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:30:13.733258 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:30:13.733270 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.772727 | |
| I0506 00:30:13.733283 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.075 | |
| I0506 00:30:13.733296 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.91285 (* 0.3 = 1.17386 loss) | |
| I0506 00:30:13.733310 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.36105 (* 0.3 = 0.408314 loss) | |
| I0506 00:30:13.733326 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.87384 (* 0.0272727 = 0.10565 loss) | |
| I0506 00:30:13.733341 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.84855 (* 0.0272727 = 0.10496 loss) | |
| I0506 00:30:13.733355 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.67644 (* 0.0272727 = 0.100267 loss) | |
| I0506 00:30:13.733381 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.86126 (* 0.0272727 = 0.105307 loss) | |
| I0506 00:30:13.733397 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.75571 (* 0.0272727 = 0.0751558 loss) | |
| I0506 00:30:13.733410 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.17223 (* 0.0272727 = 0.0592426 loss) | |
| I0506 00:30:13.733423 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.5886 (* 0.0272727 = 0.0433255 loss) | |
| I0506 00:30:13.733438 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.995935 (* 0.0272727 = 0.0271619 loss) | |
| I0506 00:30:13.733450 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.981473 (* 0.0272727 = 0.0267674 loss) | |
| I0506 00:30:13.733465 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.154178 (* 0.0272727 = 0.00420485 loss) | |
| I0506 00:30:13.733479 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.141413 (* 0.0272727 = 0.00385673 loss) | |
| I0506 00:30:13.733492 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0947652 (* 0.0272727 = 0.0025845 loss) | |
| I0506 00:30:13.733506 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0992798 (* 0.0272727 = 0.00270763 loss) | |
| I0506 00:30:13.733520 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0811393 (* 0.0272727 = 0.00221289 loss) | |
| I0506 00:30:13.733535 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0978261 (* 0.0272727 = 0.00266798 loss) | |
| I0506 00:30:13.733547 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.048119 (* 0.0272727 = 0.00131234 loss) | |
| I0506 00:30:13.733561 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0602379 (* 0.0272727 = 0.00164285 loss) | |
| I0506 00:30:13.733574 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0339662 (* 0.0272727 = 0.000926352 loss) | |
| I0506 00:30:13.733588 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0462804 (* 0.0272727 = 0.00126219 loss) | |
| I0506 00:30:13.733603 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.064329 (* 0.0272727 = 0.00175443 loss) | |
| I0506 00:30:13.733615 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0476948 (* 0.0272727 = 0.00130077 loss) | |
| I0506 00:30:13.733629 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0361015 (* 0.0272727 = 0.000984586 loss) | |
| I0506 00:30:13.733641 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.05 | |
| I0506 00:30:13.733654 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 00:30:13.733665 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 00:30:13.733676 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:30:13.733687 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0506 00:30:13.733698 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 00:30:13.733711 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 00:30:13.733722 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 00:30:13.733733 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:30:13.733746 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:30:13.733757 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:30:13.733768 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:30:13.733779 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:30:13.733791 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:30:13.733803 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:30:13.733814 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:30:13.733825 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:30:13.733836 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:30:13.733857 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:30:13.733870 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:30:13.733881 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:30:13.733892 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:30:13.733903 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:30:13.733916 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.761364 | |
| I0506 00:30:13.733929 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.15 | |
| I0506 00:30:13.733943 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.80217 (* 1 = 3.80217 loss) | |
| I0506 00:30:13.733958 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.13843 (* 1 = 1.13843 loss) | |
| I0506 00:30:13.733970 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.66058 (* 0.0909091 = 0.33278 loss) | |
| I0506 00:30:13.733985 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.53179 (* 0.0909091 = 0.321072 loss) | |
| I0506 00:30:13.733999 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.69976 (* 0.0909091 = 0.336342 loss) | |
| I0506 00:30:13.734012 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.35831 (* 0.0909091 = 0.305301 loss) | |
| I0506 00:30:13.734025 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.85152 (* 0.0909091 = 0.259229 loss) | |
| I0506 00:30:13.734040 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.25952 (* 0.0909091 = 0.205411 loss) | |
| I0506 00:30:13.734052 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.50918 (* 0.0909091 = 0.137198 loss) | |
| I0506 00:30:13.734066 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.784802 (* 0.0909091 = 0.0713456 loss) | |
| I0506 00:30:13.734079 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.815925 (* 0.0909091 = 0.074175 loss) | |
| I0506 00:30:13.734094 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0512112 (* 0.0909091 = 0.00465557 loss) | |
| I0506 00:30:13.734107 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0463037 (* 0.0909091 = 0.00420943 loss) | |
| I0506 00:30:13.734122 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0283051 (* 0.0909091 = 0.00257319 loss) | |
| I0506 00:30:13.734135 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0227992 (* 0.0909091 = 0.00207265 loss) | |
| I0506 00:30:13.734148 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.020244 (* 0.0909091 = 0.00184036 loss) | |
| I0506 00:30:13.734163 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0116791 (* 0.0909091 = 0.00106173 loss) | |
| I0506 00:30:13.734175 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00587467 (* 0.0909091 = 0.000534061 loss) | |
| I0506 00:30:13.734189 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00324904 (* 0.0909091 = 0.000295368 loss) | |
| I0506 00:30:13.734202 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00193283 (* 0.0909091 = 0.000175712 loss) | |
| I0506 00:30:13.734217 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00204906 (* 0.0909091 = 0.000186279 loss) | |
| I0506 00:30:13.734230 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00177996 (* 0.0909091 = 0.000161815 loss) | |
| I0506 00:30:13.734244 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00160649 (* 0.0909091 = 0.000146044 loss) | |
| I0506 00:30:13.734258 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00166082 (* 0.0909091 = 0.000150984 loss) | |
| I0506 00:30:13.734271 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:30:13.734282 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:30:13.734292 15760 solver.cpp:245] Train net output #149: total_confidence = 2.48768e-07 | |
| I0506 00:30:13.734314 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.7664e-06 | |
| I0506 00:30:13.734329 15760 sgd_solver.cpp:106] Iteration 2500, lr = 0.001 | |
| I0506 00:32:00.868652 15760 solver.cpp:229] Iteration 3000, loss = 12.6213 | |
| I0506 00:32:00.868798 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:32:00.868818 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:32:00.868830 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:32:00.868842 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:32:00.868854 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 00:32:00.868866 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 00:32:00.868881 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 00:32:00.868893 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:32:00.868906 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 00:32:00.868918 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:32:00.868929 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:32:00.868942 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:32:00.868954 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 00:32:00.868966 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:32:00.868978 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:32:00.868990 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:32:00.869002 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:32:00.869014 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:32:00.869026 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:32:00.869038 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:32:00.869050 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:32:00.869061 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:32:00.869073 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:32:00.869084 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0506 00:32:00.869096 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0980392 | |
| I0506 00:32:00.869113 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.84904 (* 0.3 = 1.15471 loss) | |
| I0506 00:32:00.869143 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.31798 (* 0.3 = 0.395393 loss) | |
| I0506 00:32:00.869158 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.90403 (* 0.0272727 = 0.106474 loss) | |
| I0506 00:32:00.869171 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.59418 (* 0.0272727 = 0.0980231 loss) | |
| I0506 00:32:00.869185 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.19533 (* 0.0272727 = 0.114418 loss) | |
| I0506 00:32:00.869199 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.87634 (* 0.0272727 = 0.105718 loss) | |
| I0506 00:32:00.869213 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.12591 (* 0.0272727 = 0.085252 loss) | |
| I0506 00:32:00.869227 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.24565 (* 0.0272727 = 0.0885178 loss) | |
| I0506 00:32:00.869241 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.70176 (* 0.0272727 = 0.0736844 loss) | |
| I0506 00:32:00.869254 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.43939 (* 0.0272727 = 0.0392562 loss) | |
| I0506 00:32:00.869268 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.15144 (* 0.0272727 = 0.0314028 loss) | |
| I0506 00:32:00.869282 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.93976 (* 0.0272727 = 0.0256298 loss) | |
| I0506 00:32:00.869295 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.962259 (* 0.0272727 = 0.0262434 loss) | |
| I0506 00:32:00.869309 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 1.31591 (* 0.0272727 = 0.0358883 loss) | |
| I0506 00:32:00.869323 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.120271 (* 0.0272727 = 0.00328011 loss) | |
| I0506 00:32:00.869357 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0549085 (* 0.0272727 = 0.00149751 loss) | |
| I0506 00:32:00.869374 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0327472 (* 0.0272727 = 0.000893105 loss) | |
| I0506 00:32:00.869387 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0485638 (* 0.0272727 = 0.00132447 loss) | |
| I0506 00:32:00.869401 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0356977 (* 0.0272727 = 0.000973573 loss) | |
| I0506 00:32:00.869415 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0356044 (* 0.0272727 = 0.000971029 loss) | |
| I0506 00:32:00.869429 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0351435 (* 0.0272727 = 0.000958459 loss) | |
| I0506 00:32:00.869443 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.025078 (* 0.0272727 = 0.000683945 loss) | |
| I0506 00:32:00.869457 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0217307 (* 0.0272727 = 0.000592656 loss) | |
| I0506 00:32:00.869472 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0540666 (* 0.0272727 = 0.00147454 loss) | |
| I0506 00:32:00.869484 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0196078 | |
| I0506 00:32:00.869496 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:32:00.869509 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:32:00.869520 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 00:32:00.869532 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:32:00.869544 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 00:32:00.869556 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 00:32:00.869568 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:32:00.869580 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 00:32:00.869591 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:32:00.869603 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:32:00.869616 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:32:00.869626 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 00:32:00.869638 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:32:00.869649 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:32:00.869662 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:32:00.869673 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:32:00.869683 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:32:00.869695 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:32:00.869706 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:32:00.869717 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:32:00.869729 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:32:00.869740 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:32:00.869752 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.715909 | |
| I0506 00:32:00.869763 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0588235 | |
| I0506 00:32:00.869777 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.02809 (* 0.3 = 1.20843 loss) | |
| I0506 00:32:00.869791 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.31618 (* 0.3 = 0.394853 loss) | |
| I0506 00:32:00.869806 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.59261 (* 0.0272727 = 0.0979802 loss) | |
| I0506 00:32:00.869823 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.82711 (* 0.0272727 = 0.104376 loss) | |
| I0506 00:32:00.869849 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.79382 (* 0.0272727 = 0.103468 loss) | |
| I0506 00:32:00.869863 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.50746 (* 0.0272727 = 0.0956581 loss) | |
| I0506 00:32:00.869877 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.43224 (* 0.0272727 = 0.0936066 loss) | |
| I0506 00:32:00.869891 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.05017 (* 0.0272727 = 0.0831865 loss) | |
| I0506 00:32:00.869905 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.85447 (* 0.0272727 = 0.0778493 loss) | |
| I0506 00:32:00.869918 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.58285 (* 0.0272727 = 0.0431688 loss) | |
| I0506 00:32:00.869935 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.868676 (* 0.0272727 = 0.0236912 loss) | |
| I0506 00:32:00.869949 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.973399 (* 0.0272727 = 0.0265472 loss) | |
| I0506 00:32:00.869963 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.960614 (* 0.0272727 = 0.0261986 loss) | |
| I0506 00:32:00.869977 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.957529 (* 0.0272727 = 0.0261144 loss) | |
| I0506 00:32:00.869992 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0834575 (* 0.0272727 = 0.00227611 loss) | |
| I0506 00:32:00.870004 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.104621 (* 0.0272727 = 0.00285331 loss) | |
| I0506 00:32:00.870018 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.074547 (* 0.0272727 = 0.0020331 loss) | |
| I0506 00:32:00.870033 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0563175 (* 0.0272727 = 0.00153593 loss) | |
| I0506 00:32:00.870048 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0383428 (* 0.0272727 = 0.00104571 loss) | |
| I0506 00:32:00.870061 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0426646 (* 0.0272727 = 0.00116358 loss) | |
| I0506 00:32:00.870074 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0521135 (* 0.0272727 = 0.00142128 loss) | |
| I0506 00:32:00.870088 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0372527 (* 0.0272727 = 0.00101598 loss) | |
| I0506 00:32:00.870102 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0429663 (* 0.0272727 = 0.00117181 loss) | |
| I0506 00:32:00.870117 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0331418 (* 0.0272727 = 0.000903868 loss) | |
| I0506 00:32:00.870129 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0506 00:32:00.870141 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:32:00.870152 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:32:00.870163 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:32:00.870175 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:32:00.870187 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 00:32:00.870198 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 00:32:00.870209 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:32:00.870221 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 00:32:00.870232 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:32:00.870244 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:32:00.870255 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:32:00.870267 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 00:32:00.870280 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:32:00.870290 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:32:00.870301 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:32:00.870313 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:32:00.870334 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:32:00.870347 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:32:00.870358 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:32:00.870370 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:32:00.870381 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:32:00.870393 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:32:00.870404 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0506 00:32:00.870416 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.176471 | |
| I0506 00:32:00.870430 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.72663 (* 1 = 3.72663 loss) | |
| I0506 00:32:00.870443 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.24754 (* 1 = 1.24754 loss) | |
| I0506 00:32:00.870457 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.58997 (* 0.0909091 = 0.326361 loss) | |
| I0506 00:32:00.870471 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.62278 (* 0.0909091 = 0.329344 loss) | |
| I0506 00:32:00.870484 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.68129 (* 0.0909091 = 0.334663 loss) | |
| I0506 00:32:00.870498 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.34572 (* 0.0909091 = 0.304156 loss) | |
| I0506 00:32:00.870512 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.89505 (* 0.0909091 = 0.263186 loss) | |
| I0506 00:32:00.870525 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.81631 (* 0.0909091 = 0.256028 loss) | |
| I0506 00:32:00.870538 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.56669 (* 0.0909091 = 0.233336 loss) | |
| I0506 00:32:00.870553 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.23629 (* 0.0909091 = 0.11239 loss) | |
| I0506 00:32:00.870565 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.925374 (* 0.0909091 = 0.0841249 loss) | |
| I0506 00:32:00.870579 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.757541 (* 0.0909091 = 0.0688674 loss) | |
| I0506 00:32:00.870592 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.683039 (* 0.0909091 = 0.0620945 loss) | |
| I0506 00:32:00.870606 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.786719 (* 0.0909091 = 0.0715199 loss) | |
| I0506 00:32:00.870620 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0839267 (* 0.0909091 = 0.0076297 loss) | |
| I0506 00:32:00.870633 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0585565 (* 0.0909091 = 0.00532331 loss) | |
| I0506 00:32:00.870647 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0358712 (* 0.0909091 = 0.00326102 loss) | |
| I0506 00:32:00.870661 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.028012 (* 0.0909091 = 0.00254654 loss) | |
| I0506 00:32:00.870674 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0201999 (* 0.0909091 = 0.00183635 loss) | |
| I0506 00:32:00.870688 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0173667 (* 0.0909091 = 0.00157879 loss) | |
| I0506 00:32:00.870702 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0127893 (* 0.0909091 = 0.00116266 loss) | |
| I0506 00:32:00.870715 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0111715 (* 0.0909091 = 0.00101559 loss) | |
| I0506 00:32:00.870729 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.013716 (* 0.0909091 = 0.00124691 loss) | |
| I0506 00:32:00.870743 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00916522 (* 0.0909091 = 0.000833202 loss) | |
| I0506 00:32:00.870755 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:32:00.870766 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:32:00.870777 15760 solver.cpp:245] Train net output #149: total_confidence = 2.28998e-07 | |
| I0506 00:32:00.870798 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.81395e-06 | |
| I0506 00:32:00.870813 15760 sgd_solver.cpp:106] Iteration 3000, lr = 0.001 | |
| I0506 00:33:47.828197 15760 solver.cpp:229] Iteration 3500, loss = 12.3009 | |
| I0506 00:33:47.828315 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:33:47.828335 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:33:47.828346 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:33:47.828358 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:33:47.828371 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:33:47.828382 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 00:33:47.828393 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 00:33:47.828405 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 00:33:47.828418 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:33:47.828429 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:33:47.828441 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:33:47.828454 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:33:47.828464 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 00:33:47.828476 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 00:33:47.828487 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:33:47.828500 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:33:47.828512 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:33:47.828534 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:33:47.828548 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:33:47.828559 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:33:47.828570 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:33:47.828583 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:33:47.828593 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:33:47.828604 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.727273 | |
| I0506 00:33:47.828616 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0625 | |
| I0506 00:33:47.828632 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.03322 (* 0.3 = 1.20997 loss) | |
| I0506 00:33:47.828646 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.43338 (* 0.3 = 0.430013 loss) | |
| I0506 00:33:47.828660 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.57503 (* 0.0272727 = 0.0975008 loss) | |
| I0506 00:33:47.828675 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.82702 (* 0.0272727 = 0.104373 loss) | |
| I0506 00:33:47.828687 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.82907 (* 0.0272727 = 0.104429 loss) | |
| I0506 00:33:47.828701 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.85485 (* 0.0272727 = 0.105132 loss) | |
| I0506 00:33:47.828714 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.05619 (* 0.0272727 = 0.0833507 loss) | |
| I0506 00:33:47.828728 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.16594 (* 0.0272727 = 0.059071 loss) | |
| I0506 00:33:47.828742 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.21279 (* 0.0272727 = 0.0330761 loss) | |
| I0506 00:33:47.828755 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.807349 (* 0.0272727 = 0.0220186 loss) | |
| I0506 00:33:47.828768 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.833934 (* 0.0272727 = 0.0227437 loss) | |
| I0506 00:33:47.828783 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.833415 (* 0.0272727 = 0.0227295 loss) | |
| I0506 00:33:47.828796 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.701894 (* 0.0272727 = 0.0191426 loss) | |
| I0506 00:33:47.828810 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.905934 (* 0.0272727 = 0.0247073 loss) | |
| I0506 00:33:47.828824 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.739747 (* 0.0272727 = 0.0201749 loss) | |
| I0506 00:33:47.828860 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0809727 (* 0.0272727 = 0.00220835 loss) | |
| I0506 00:33:47.828876 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0453824 (* 0.0272727 = 0.0012377 loss) | |
| I0506 00:33:47.828889 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0426141 (* 0.0272727 = 0.0011622 loss) | |
| I0506 00:33:47.828903 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0601587 (* 0.0272727 = 0.00164069 loss) | |
| I0506 00:33:47.828919 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0299801 (* 0.0272727 = 0.00081764 loss) | |
| I0506 00:33:47.828943 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0285102 (* 0.0272727 = 0.00077755 loss) | |
| I0506 00:33:47.828958 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0315874 (* 0.0272727 = 0.000861475 loss) | |
| I0506 00:33:47.828971 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0341306 (* 0.0272727 = 0.000930835 loss) | |
| I0506 00:33:47.828985 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0356923 (* 0.0272727 = 0.000973427 loss) | |
| I0506 00:33:47.828997 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:33:47.829010 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:33:47.829020 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:33:47.829032 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 00:33:47.829043 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 00:33:47.829054 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 00:33:47.829066 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 00:33:47.829077 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 00:33:47.829089 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:33:47.829114 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:33:47.829130 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:33:47.829143 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:33:47.829154 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 00:33:47.829164 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 00:33:47.829176 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:33:47.829187 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:33:47.829198 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:33:47.829210 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:33:47.829221 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:33:47.829232 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:33:47.829243 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:33:47.829254 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:33:47.829265 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:33:47.829277 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0506 00:33:47.829288 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.104167 | |
| I0506 00:33:47.829301 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.11886 (* 0.3 = 1.23566 loss) | |
| I0506 00:33:47.829315 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.46064 (* 0.3 = 0.438191 loss) | |
| I0506 00:33:47.829329 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.84576 (* 0.0272727 = 0.104884 loss) | |
| I0506 00:33:47.829342 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.81715 (* 0.0272727 = 0.104104 loss) | |
| I0506 00:33:47.829368 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.40583 (* 0.0272727 = 0.0928861 loss) | |
| I0506 00:33:47.829383 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.81701 (* 0.0272727 = 0.1041 loss) | |
| I0506 00:33:47.829397 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.88249 (* 0.0272727 = 0.0786133 loss) | |
| I0506 00:33:47.829411 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.29229 (* 0.0272727 = 0.0625169 loss) | |
| I0506 00:33:47.829424 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.42444 (* 0.0272727 = 0.0388483 loss) | |
| I0506 00:33:47.829437 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.704689 (* 0.0272727 = 0.0192188 loss) | |
| I0506 00:33:47.829450 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.699383 (* 0.0272727 = 0.0190741 loss) | |
| I0506 00:33:47.829464 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.683791 (* 0.0272727 = 0.0186488 loss) | |
| I0506 00:33:47.829478 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.818273 (* 0.0272727 = 0.0223165 loss) | |
| I0506 00:33:47.829495 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.632378 (* 0.0272727 = 0.0172467 loss) | |
| I0506 00:33:47.829522 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.661038 (* 0.0272727 = 0.0180283 loss) | |
| I0506 00:33:47.829541 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0930086 (* 0.0272727 = 0.0025366 loss) | |
| I0506 00:33:47.829556 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0516917 (* 0.0272727 = 0.00140977 loss) | |
| I0506 00:33:47.829571 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0634086 (* 0.0272727 = 0.00172933 loss) | |
| I0506 00:33:47.829584 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0313486 (* 0.0272727 = 0.000854961 loss) | |
| I0506 00:33:47.829597 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0220616 (* 0.0272727 = 0.000601679 loss) | |
| I0506 00:33:47.829612 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.046253 (* 0.0272727 = 0.00126145 loss) | |
| I0506 00:33:47.829625 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0208509 (* 0.0272727 = 0.00056866 loss) | |
| I0506 00:33:47.829639 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0347184 (* 0.0272727 = 0.000946866 loss) | |
| I0506 00:33:47.829655 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0293512 (* 0.0272727 = 0.000800487 loss) | |
| I0506 00:33:47.829679 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0208333 | |
| I0506 00:33:47.829695 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:33:47.829707 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:33:47.829718 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:33:47.829730 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 00:33:47.829741 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 00:33:47.829752 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 00:33:47.829764 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 00:33:47.829776 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:33:47.829787 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:33:47.829798 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:33:47.829809 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:33:47.829820 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 00:33:47.829833 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 00:33:47.829843 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:33:47.829855 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:33:47.829866 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:33:47.829887 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:33:47.829900 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:33:47.829916 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:33:47.829926 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:33:47.829937 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:33:47.829949 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:33:47.829960 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.710227 | |
| I0506 00:33:47.829972 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.125 | |
| I0506 00:33:47.829984 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.87512 (* 1 = 3.87512 loss) | |
| I0506 00:33:47.829998 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.36801 (* 1 = 1.36801 loss) | |
| I0506 00:33:47.830011 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.68714 (* 0.0909091 = 0.335195 loss) | |
| I0506 00:33:47.830025 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.4805 (* 0.0909091 = 0.316409 loss) | |
| I0506 00:33:47.830039 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.82876 (* 0.0909091 = 0.348069 loss) | |
| I0506 00:33:47.830051 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.76502 (* 0.0909091 = 0.342275 loss) | |
| I0506 00:33:47.830065 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.01714 (* 0.0909091 = 0.274285 loss) | |
| I0506 00:33:47.830078 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.35207 (* 0.0909091 = 0.213824 loss) | |
| I0506 00:33:47.830091 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.40258 (* 0.0909091 = 0.127507 loss) | |
| I0506 00:33:47.830106 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.546795 (* 0.0909091 = 0.0497087 loss) | |
| I0506 00:33:47.830118 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.623232 (* 0.0909091 = 0.0566575 loss) | |
| I0506 00:33:47.830132 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.593915 (* 0.0909091 = 0.0539922 loss) | |
| I0506 00:33:47.830145 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.672493 (* 0.0909091 = 0.0611357 loss) | |
| I0506 00:33:47.830163 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.576907 (* 0.0909091 = 0.0524461 loss) | |
| I0506 00:33:47.830178 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.684968 (* 0.0909091 = 0.0622699 loss) | |
| I0506 00:33:47.830188 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0262888 (* 0.0909091 = 0.00238989 loss) | |
| I0506 00:33:47.830204 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0241652 (* 0.0909091 = 0.00219683 loss) | |
| I0506 00:33:47.830224 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0161979 (* 0.0909091 = 0.00147254 loss) | |
| I0506 00:33:47.830245 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00708229 (* 0.0909091 = 0.000643845 loss) | |
| I0506 00:33:47.830258 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00532999 (* 0.0909091 = 0.000484545 loss) | |
| I0506 00:33:47.830271 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00583906 (* 0.0909091 = 0.000530824 loss) | |
| I0506 00:33:47.830286 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00496013 (* 0.0909091 = 0.000450921 loss) | |
| I0506 00:33:47.830299 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00555014 (* 0.0909091 = 0.000504558 loss) | |
| I0506 00:33:47.830312 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00337735 (* 0.0909091 = 0.000307032 loss) | |
| I0506 00:33:47.830324 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:33:47.830335 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:33:47.830346 15760 solver.cpp:245] Train net output #149: total_confidence = 8.86271e-08 | |
| I0506 00:33:47.830368 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 4.65825e-06 | |
| I0506 00:33:47.830382 15760 sgd_solver.cpp:106] Iteration 3500, lr = 0.001 | |
| I0506 00:35:34.941033 15760 solver.cpp:229] Iteration 4000, loss = 12.254 | |
| I0506 00:35:34.941159 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:35:34.941176 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:35:34.941190 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 00:35:34.941201 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 00:35:34.941213 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 00:35:34.941226 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0506 00:35:34.941236 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 00:35:34.941248 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 00:35:34.941260 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 00:35:34.941275 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:35:34.941287 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:35:34.941299 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:35:34.941310 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:35:34.941321 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:35:34.941334 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:35:34.941344 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:35:34.941355 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:35:34.941367 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:35:34.941378 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:35:34.941390 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:35:34.941401 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:35:34.941412 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:35:34.941423 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:35:34.941434 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.778409 | |
| I0506 00:35:34.941447 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0769231 | |
| I0506 00:35:34.941470 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.07275 (* 0.3 = 1.22183 loss) | |
| I0506 00:35:34.941485 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.17727 (* 0.3 = 0.353182 loss) | |
| I0506 00:35:34.941499 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.41745 (* 0.0272727 = 0.120476 loss) | |
| I0506 00:35:34.941514 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.74258 (* 0.0272727 = 0.10207 loss) | |
| I0506 00:35:34.941526 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.63972 (* 0.0272727 = 0.099265 loss) | |
| I0506 00:35:34.941540 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 4.2451 (* 0.0272727 = 0.115775 loss) | |
| I0506 00:35:34.941553 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.76502 (* 0.0272727 = 0.0754097 loss) | |
| I0506 00:35:34.941567 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.1064 (* 0.0272727 = 0.0574473 loss) | |
| I0506 00:35:34.941581 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.775062 (* 0.0272727 = 0.0211381 loss) | |
| I0506 00:35:34.941594 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.209599 (* 0.0272727 = 0.00571634 loss) | |
| I0506 00:35:34.941608 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.178712 (* 0.0272727 = 0.00487397 loss) | |
| I0506 00:35:34.941622 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0996949 (* 0.0272727 = 0.00271895 loss) | |
| I0506 00:35:34.941635 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0834083 (* 0.0272727 = 0.00227477 loss) | |
| I0506 00:35:34.941649 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0612707 (* 0.0272727 = 0.00167102 loss) | |
| I0506 00:35:34.941663 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0505425 (* 0.0272727 = 0.00137843 loss) | |
| I0506 00:35:34.941695 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0714976 (* 0.0272727 = 0.00194993 loss) | |
| I0506 00:35:34.941710 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0460242 (* 0.0272727 = 0.00125521 loss) | |
| I0506 00:35:34.941725 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0419106 (* 0.0272727 = 0.00114301 loss) | |
| I0506 00:35:34.941738 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0173208 (* 0.0272727 = 0.000472385 loss) | |
| I0506 00:35:34.941751 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0201159 (* 0.0272727 = 0.000548615 loss) | |
| I0506 00:35:34.941766 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0226381 (* 0.0272727 = 0.000617404 loss) | |
| I0506 00:35:34.941779 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0135586 (* 0.0272727 = 0.00036978 loss) | |
| I0506 00:35:34.941793 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0166058 (* 0.0272727 = 0.000452886 loss) | |
| I0506 00:35:34.941807 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0187883 (* 0.0272727 = 0.000512407 loss) | |
| I0506 00:35:34.941818 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:35:34.941830 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:35:34.941841 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:35:34.941853 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:35:34.941864 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:35:34.941874 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 00:35:34.941890 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 00:35:34.941901 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 00:35:34.941913 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 00:35:34.941925 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:35:34.941936 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:35:34.941946 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:35:34.941958 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:35:34.941965 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:35:34.941972 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:35:34.941984 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:35:34.941997 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:35:34.942008 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:35:34.942018 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:35:34.942035 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:35:34.942047 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:35:34.942059 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:35:34.942070 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:35:34.942080 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0506 00:35:34.942092 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.128205 | |
| I0506 00:35:34.942106 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.99793 (* 0.3 = 1.19938 loss) | |
| I0506 00:35:34.942119 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.22247 (* 0.3 = 0.366741 loss) | |
| I0506 00:35:34.942133 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.13107 (* 0.0272727 = 0.112666 loss) | |
| I0506 00:35:34.942147 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.48123 (* 0.0272727 = 0.0949426 loss) | |
| I0506 00:35:34.942160 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 4.18514 (* 0.0272727 = 0.11414 loss) | |
| I0506 00:35:34.942185 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.84139 (* 0.0272727 = 0.104765 loss) | |
| I0506 00:35:34.942200 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.09699 (* 0.0272727 = 0.0844634 loss) | |
| I0506 00:35:34.942214 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.26481 (* 0.0272727 = 0.0617675 loss) | |
| I0506 00:35:34.942227 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.837699 (* 0.0272727 = 0.0228463 loss) | |
| I0506 00:35:34.942241 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.25565 (* 0.0272727 = 0.00697226 loss) | |
| I0506 00:35:34.942255 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.159062 (* 0.0272727 = 0.00433806 loss) | |
| I0506 00:35:34.942268 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.114289 (* 0.0272727 = 0.00311697 loss) | |
| I0506 00:35:34.942282 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.14157 (* 0.0272727 = 0.00386101 loss) | |
| I0506 00:35:34.942296 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.122838 (* 0.0272727 = 0.00335011 loss) | |
| I0506 00:35:34.942309 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.114564 (* 0.0272727 = 0.00312448 loss) | |
| I0506 00:35:34.942327 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.118745 (* 0.0272727 = 0.00323849 loss) | |
| I0506 00:35:34.942342 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.086947 (* 0.0272727 = 0.00237128 loss) | |
| I0506 00:35:34.942355 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0545653 (* 0.0272727 = 0.00148814 loss) | |
| I0506 00:35:34.942369 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0568515 (* 0.0272727 = 0.00155049 loss) | |
| I0506 00:35:34.942383 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0425192 (* 0.0272727 = 0.00115961 loss) | |
| I0506 00:35:34.942395 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0560162 (* 0.0272727 = 0.00152771 loss) | |
| I0506 00:35:34.942409 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0476825 (* 0.0272727 = 0.00130043 loss) | |
| I0506 00:35:34.942422 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0433958 (* 0.0272727 = 0.00118352 loss) | |
| I0506 00:35:34.942436 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0276074 (* 0.0272727 = 0.000752928 loss) | |
| I0506 00:35:34.942447 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.128205 | |
| I0506 00:35:34.942459 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 00:35:34.942471 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:35:34.942482 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 00:35:34.942493 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:35:34.942504 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 00:35:34.942515 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 00:35:34.942528 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0506 00:35:34.942538 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 00:35:34.942549 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:35:34.942560 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:35:34.942571 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:35:34.942582 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:35:34.942594 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:35:34.942605 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:35:34.942616 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:35:34.942627 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:35:34.942637 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:35:34.942659 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:35:34.942672 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:35:34.942683 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:35:34.942694 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:35:34.942705 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:35:34.942716 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.806818 | |
| I0506 00:35:34.942728 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.230769 | |
| I0506 00:35:34.942741 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.6056 (* 1 = 3.6056 loss) | |
| I0506 00:35:34.942755 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.935333 (* 1 = 0.935333 loss) | |
| I0506 00:35:34.942769 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 4.11263 (* 0.0909091 = 0.373875 loss) | |
| I0506 00:35:34.942782 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.42022 (* 0.0909091 = 0.310929 loss) | |
| I0506 00:35:34.942795 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.79834 (* 0.0909091 = 0.345304 loss) | |
| I0506 00:35:34.942808 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.88476 (* 0.0909091 = 0.35316 loss) | |
| I0506 00:35:34.942822 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.7687 (* 0.0909091 = 0.2517 loss) | |
| I0506 00:35:34.942836 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.07317 (* 0.0909091 = 0.18847 loss) | |
| I0506 00:35:34.942848 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.757425 (* 0.0909091 = 0.0688568 loss) | |
| I0506 00:35:34.942862 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.193482 (* 0.0909091 = 0.0175893 loss) | |
| I0506 00:35:34.942876 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.111617 (* 0.0909091 = 0.010147 loss) | |
| I0506 00:35:34.942889 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.117132 (* 0.0909091 = 0.0106483 loss) | |
| I0506 00:35:34.942903 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0791824 (* 0.0909091 = 0.0071984 loss) | |
| I0506 00:35:34.942916 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0581582 (* 0.0909091 = 0.00528711 loss) | |
| I0506 00:35:34.942930 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0470163 (* 0.0909091 = 0.00427421 loss) | |
| I0506 00:35:34.942947 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.044675 (* 0.0909091 = 0.00406136 loss) | |
| I0506 00:35:34.942961 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0306138 (* 0.0909091 = 0.00278307 loss) | |
| I0506 00:35:34.942975 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0135754 (* 0.0909091 = 0.00123413 loss) | |
| I0506 00:35:34.942988 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0104848 (* 0.0909091 = 0.000953165 loss) | |
| I0506 00:35:34.943002 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00659729 (* 0.0909091 = 0.000599753 loss) | |
| I0506 00:35:34.943016 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00546778 (* 0.0909091 = 0.000497071 loss) | |
| I0506 00:35:34.943029 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00344047 (* 0.0909091 = 0.00031277 loss) | |
| I0506 00:35:34.943042 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00277346 (* 0.0909091 = 0.000252133 loss) | |
| I0506 00:35:34.943056 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0026065 (* 0.0909091 = 0.000236955 loss) | |
| I0506 00:35:34.943068 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:35:34.943079 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:35:34.943090 15760 solver.cpp:245] Train net output #149: total_confidence = 1.56474e-05 | |
| I0506 00:35:34.943101 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 3.18569e-05 | |
| I0506 00:35:34.943123 15760 sgd_solver.cpp:106] Iteration 4000, lr = 0.001 | |
| I0506 00:37:22.028241 15760 solver.cpp:229] Iteration 4500, loss = 12.0236 | |
| I0506 00:37:22.028367 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0925926 | |
| I0506 00:37:22.028388 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:37:22.028400 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:37:22.028411 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:37:22.028424 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:37:22.028434 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:37:22.028446 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0506 00:37:22.028458 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0506 00:37:22.028470 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.5 | |
| I0506 00:37:22.028481 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:37:22.028493 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:37:22.028504 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:37:22.028517 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:37:22.028527 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:37:22.028538 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:37:22.028556 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:37:22.028574 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:37:22.028587 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:37:22.028599 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:37:22.028610 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:37:22.028621 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:37:22.028632 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:37:22.028643 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:37:22.028656 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0506 00:37:22.028666 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.222222 | |
| I0506 00:37:22.028682 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.59925 (* 0.3 = 1.07978 loss) | |
| I0506 00:37:22.028697 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.33753 (* 0.3 = 0.40126 loss) | |
| I0506 00:37:22.028710 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.74171 (* 0.0272727 = 0.102047 loss) | |
| I0506 00:37:22.028724 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.39782 (* 0.0272727 = 0.0926677 loss) | |
| I0506 00:37:22.028738 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.8293 (* 0.0272727 = 0.104435 loss) | |
| I0506 00:37:22.028751 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.63609 (* 0.0272727 = 0.099166 loss) | |
| I0506 00:37:22.028765 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.85098 (* 0.0272727 = 0.105027 loss) | |
| I0506 00:37:22.028779 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.39209 (* 0.0272727 = 0.0925116 loss) | |
| I0506 00:37:22.028792 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 3.30596 (* 0.0272727 = 0.0901625 loss) | |
| I0506 00:37:22.028806 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 2.7968 (* 0.0272727 = 0.0762763 loss) | |
| I0506 00:37:22.028821 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.104119 (* 0.0272727 = 0.00283961 loss) | |
| I0506 00:37:22.028836 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0947401 (* 0.0272727 = 0.00258382 loss) | |
| I0506 00:37:22.028858 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.110597 (* 0.0272727 = 0.00301627 loss) | |
| I0506 00:37:22.028877 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0599865 (* 0.0272727 = 0.00163599 loss) | |
| I0506 00:37:22.028893 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0373423 (* 0.0272727 = 0.00101843 loss) | |
| I0506 00:37:22.028924 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.027426 (* 0.0272727 = 0.000747981 loss) | |
| I0506 00:37:22.028940 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0323856 (* 0.0272727 = 0.000883242 loss) | |
| I0506 00:37:22.028954 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0220958 (* 0.0272727 = 0.000602612 loss) | |
| I0506 00:37:22.028969 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0212059 (* 0.0272727 = 0.000578342 loss) | |
| I0506 00:37:22.028982 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0129363 (* 0.0272727 = 0.000352809 loss) | |
| I0506 00:37:22.028995 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0120446 (* 0.0272727 = 0.00032849 loss) | |
| I0506 00:37:22.029009 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0125592 (* 0.0272727 = 0.000342525 loss) | |
| I0506 00:37:22.029023 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0154013 (* 0.0272727 = 0.000420037 loss) | |
| I0506 00:37:22.029037 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0141938 (* 0.0272727 = 0.000387104 loss) | |
| I0506 00:37:22.029049 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.037037 | |
| I0506 00:37:22.029062 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:37:22.029073 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:37:22.029084 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 00:37:22.029096 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 00:37:22.029108 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 00:37:22.029134 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0506 00:37:22.029148 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0506 00:37:22.029160 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.5 | |
| I0506 00:37:22.029171 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:37:22.029182 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:37:22.029194 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:37:22.029206 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:37:22.029217 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:37:22.029227 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:37:22.029238 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:37:22.029249 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:37:22.029260 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:37:22.029271 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:37:22.029283 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:37:22.029294 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:37:22.029305 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:37:22.029316 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:37:22.029328 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.698864 | |
| I0506 00:37:22.029340 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.203704 | |
| I0506 00:37:22.029353 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.69776 (* 0.3 = 1.10933 loss) | |
| I0506 00:37:22.029367 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.49093 (* 0.3 = 0.447278 loss) | |
| I0506 00:37:22.029384 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.51469 (* 0.0272727 = 0.0958551 loss) | |
| I0506 00:37:22.029399 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.58438 (* 0.0272727 = 0.0977559 loss) | |
| I0506 00:37:22.029413 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.37438 (* 0.0272727 = 0.0920286 loss) | |
| I0506 00:37:22.029439 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.45409 (* 0.0272727 = 0.0942025 loss) | |
| I0506 00:37:22.029454 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.2429 (* 0.0272727 = 0.0884427 loss) | |
| I0506 00:37:22.029469 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.72559 (* 0.0272727 = 0.101607 loss) | |
| I0506 00:37:22.029481 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 3.43158 (* 0.0272727 = 0.0935885 loss) | |
| I0506 00:37:22.029495 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 2.95689 (* 0.0272727 = 0.0806426 loss) | |
| I0506 00:37:22.029508 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.195549 (* 0.0272727 = 0.00533315 loss) | |
| I0506 00:37:22.029522 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.159679 (* 0.0272727 = 0.00435489 loss) | |
| I0506 00:37:22.029536 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.220008 (* 0.0272727 = 0.00600021 loss) | |
| I0506 00:37:22.029549 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.116634 (* 0.0272727 = 0.00318093 loss) | |
| I0506 00:37:22.029563 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.165492 (* 0.0272727 = 0.00451342 loss) | |
| I0506 00:37:22.029577 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.134888 (* 0.0272727 = 0.00367875 loss) | |
| I0506 00:37:22.029590 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.107014 (* 0.0272727 = 0.00291857 loss) | |
| I0506 00:37:22.029604 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.118491 (* 0.0272727 = 0.00323158 loss) | |
| I0506 00:37:22.029618 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0990855 (* 0.0272727 = 0.00270233 loss) | |
| I0506 00:37:22.029631 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0800525 (* 0.0272727 = 0.00218325 loss) | |
| I0506 00:37:22.029644 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0804018 (* 0.0272727 = 0.00219278 loss) | |
| I0506 00:37:22.029659 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0537953 (* 0.0272727 = 0.00146714 loss) | |
| I0506 00:37:22.029672 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0602264 (* 0.0272727 = 0.00164254 loss) | |
| I0506 00:37:22.029685 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0400085 (* 0.0272727 = 0.00109114 loss) | |
| I0506 00:37:22.029697 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0555556 | |
| I0506 00:37:22.029709 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:37:22.029721 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:37:22.029732 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 00:37:22.029743 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:37:22.029755 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 00:37:22.029767 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 00:37:22.029778 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0506 00:37:22.029788 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.5 | |
| I0506 00:37:22.029800 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:37:22.029811 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:37:22.029826 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:37:22.029839 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:37:22.029850 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:37:22.029871 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:37:22.029884 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:37:22.029896 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:37:22.029906 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:37:22.029932 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:37:22.029945 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:37:22.029956 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:37:22.029968 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:37:22.029978 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:37:22.029990 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0506 00:37:22.030001 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.203704 | |
| I0506 00:37:22.030014 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.33715 (* 1 = 3.33715 loss) | |
| I0506 00:37:22.030028 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.25807 (* 1 = 1.25807 loss) | |
| I0506 00:37:22.030041 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.45987 (* 0.0909091 = 0.314533 loss) | |
| I0506 00:37:22.030055 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.0107 (* 0.0909091 = 0.2737 loss) | |
| I0506 00:37:22.030068 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.29278 (* 0.0909091 = 0.299344 loss) | |
| I0506 00:37:22.030081 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.52 (* 0.0909091 = 0.32 loss) | |
| I0506 00:37:22.030094 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.08121 (* 0.0909091 = 0.28011 loss) | |
| I0506 00:37:22.030108 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.37688 (* 0.0909091 = 0.306989 loss) | |
| I0506 00:37:22.030122 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.92882 (* 0.0909091 = 0.266256 loss) | |
| I0506 00:37:22.030134 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 2.53484 (* 0.0909091 = 0.23044 loss) | |
| I0506 00:37:22.030148 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.261933 (* 0.0909091 = 0.0238121 loss) | |
| I0506 00:37:22.030160 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.148355 (* 0.0909091 = 0.0134868 loss) | |
| I0506 00:37:22.030174 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.140901 (* 0.0909091 = 0.0128091 loss) | |
| I0506 00:37:22.030187 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0798702 (* 0.0909091 = 0.00726093 loss) | |
| I0506 00:37:22.030201 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0587203 (* 0.0909091 = 0.00533821 loss) | |
| I0506 00:37:22.030215 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0351203 (* 0.0909091 = 0.00319275 loss) | |
| I0506 00:37:22.030228 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0237311 (* 0.0909091 = 0.00215737 loss) | |
| I0506 00:37:22.030242 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.011033 (* 0.0909091 = 0.001003 loss) | |
| I0506 00:37:22.030256 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00468164 (* 0.0909091 = 0.000425604 loss) | |
| I0506 00:37:22.030269 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00243345 (* 0.0909091 = 0.000221222 loss) | |
| I0506 00:37:22.030282 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00160373 (* 0.0909091 = 0.000145794 loss) | |
| I0506 00:37:22.030297 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00167546 (* 0.0909091 = 0.000152315 loss) | |
| I0506 00:37:22.030310 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00076029 (* 0.0909091 = 6.91172e-05 loss) | |
| I0506 00:37:22.030324 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000632408 (* 0.0909091 = 5.74916e-05 loss) | |
| I0506 00:37:22.030336 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:37:22.030347 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:37:22.030359 15760 solver.cpp:245] Train net output #149: total_confidence = 5.72235e-07 | |
| I0506 00:37:22.030370 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.71345e-06 | |
| I0506 00:37:22.030393 15760 sgd_solver.cpp:106] Iteration 4500, lr = 0.001 | |
| I0506 00:39:05.517979 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.2561 > 30) by scale factor 0.959812 | |
| I0506 00:39:08.933492 15760 solver.cpp:338] Iteration 5000, Testing net (#0) | |
| I0506 00:39:45.399798 15760 solver.cpp:393] Test loss: 11.2564 | |
| I0506 00:39:45.399894 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0268623 | |
| I0506 00:39:45.399914 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.06 | |
| I0506 00:39:45.399927 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.059 | |
| I0506 00:39:45.399940 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.04 | |
| I0506 00:39:45.399950 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.132 | |
| I0506 00:39:45.399962 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.287 | |
| I0506 00:39:45.399974 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.446 | |
| I0506 00:39:45.399986 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.729 | |
| I0506 00:39:45.399996 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.91 | |
| I0506 00:39:45.400012 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.99 | |
| I0506 00:39:45.400032 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0506 00:39:45.400053 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0506 00:39:45.400068 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0506 00:39:45.400079 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0506 00:39:45.400089 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0506 00:39:45.400101 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0506 00:39:45.400112 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0506 00:39:45.400123 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0506 00:39:45.400135 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0506 00:39:45.400146 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0506 00:39:45.400156 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0506 00:39:45.400167 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 00:39:45.400178 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 00:39:45.400189 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.75541 | |
| I0506 00:39:45.400202 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.116806 | |
| I0506 00:39:45.400218 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 4.54395 (* 0.3 = 1.36318 loss) | |
| I0506 00:39:45.400233 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.19011 (* 0.3 = 0.357032 loss) | |
| I0506 00:39:45.400246 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.71166 (* 0.0272727 = 0.101227 loss) | |
| I0506 00:39:45.400260 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.72333 (* 0.0272727 = 0.101545 loss) | |
| I0506 00:39:45.400274 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.79118 (* 0.0272727 = 0.103396 loss) | |
| I0506 00:39:45.400286 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.63281 (* 0.0272727 = 0.0990765 loss) | |
| I0506 00:39:45.400300 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 3.26002 (* 0.0272727 = 0.0889097 loss) | |
| I0506 00:39:45.400313 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.76716 (* 0.0272727 = 0.0754681 loss) | |
| I0506 00:39:45.400327 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.65351 (* 0.0272727 = 0.0450957 loss) | |
| I0506 00:39:45.400341 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 0.687023 (* 0.0272727 = 0.018737 loss) | |
| I0506 00:39:45.400354 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.144178 (* 0.0272727 = 0.00393212 loss) | |
| I0506 00:39:45.400368 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0926076 (* 0.0272727 = 0.00252566 loss) | |
| I0506 00:39:45.400382 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0682623 (* 0.0272727 = 0.0018617 loss) | |
| I0506 00:39:45.400395 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.059141 (* 0.0272727 = 0.00161294 loss) | |
| I0506 00:39:45.400409 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0473945 (* 0.0272727 = 0.00129258 loss) | |
| I0506 00:39:45.400442 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0463171 (* 0.0272727 = 0.00126319 loss) | |
| I0506 00:39:45.400457 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0362911 (* 0.0272727 = 0.000989756 loss) | |
| I0506 00:39:45.400471 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.0266797 (* 0.0272727 = 0.000727628 loss) | |
| I0506 00:39:45.400485 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0210556 (* 0.0272727 = 0.000574242 loss) | |
| I0506 00:39:45.400498 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0180348 (* 0.0272727 = 0.000491858 loss) | |
| I0506 00:39:45.400511 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0173178 (* 0.0272727 = 0.000472303 loss) | |
| I0506 00:39:45.400526 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0181812 (* 0.0272727 = 0.000495852 loss) | |
| I0506 00:39:45.400538 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0161144 (* 0.0272727 = 0.000439483 loss) | |
| I0506 00:39:45.400552 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0163444 (* 0.0272727 = 0.000445757 loss) | |
| I0506 00:39:45.400563 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0239448 | |
| I0506 00:39:45.400575 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.06 | |
| I0506 00:39:45.400586 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.045 | |
| I0506 00:39:45.400598 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.05 | |
| I0506 00:39:45.400609 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.133 | |
| I0506 00:39:45.400620 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.286 | |
| I0506 00:39:45.400631 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.446 | |
| I0506 00:39:45.400642 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.729 | |
| I0506 00:39:45.400653 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.91 | |
| I0506 00:39:45.400665 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.99 | |
| I0506 00:39:45.400679 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0506 00:39:45.400692 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0506 00:39:45.400703 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0506 00:39:45.400714 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0506 00:39:45.400725 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0506 00:39:45.400737 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0506 00:39:45.400748 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0506 00:39:45.400758 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0506 00:39:45.400769 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0506 00:39:45.400780 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0506 00:39:45.400791 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0506 00:39:45.400802 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 00:39:45.400813 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 00:39:45.400825 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.754638 | |
| I0506 00:39:45.400835 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.122148 | |
| I0506 00:39:45.400849 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 4.53591 (* 0.3 = 1.36077 loss) | |
| I0506 00:39:45.400863 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.18647 (* 0.3 = 0.355942 loss) | |
| I0506 00:39:45.400876 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 3.70351 (* 0.0272727 = 0.101005 loss) | |
| I0506 00:39:45.400890 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.72907 (* 0.0272727 = 0.101702 loss) | |
| I0506 00:39:45.400903 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.77849 (* 0.0272727 = 0.10305 loss) | |
| I0506 00:39:45.400930 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.63279 (* 0.0272727 = 0.0990761 loss) | |
| I0506 00:39:45.400945 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 3.27139 (* 0.0272727 = 0.0892196 loss) | |
| I0506 00:39:45.400959 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.76122 (* 0.0272727 = 0.0753059 loss) | |
| I0506 00:39:45.400972 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.65289 (* 0.0272727 = 0.0450789 loss) | |
| I0506 00:39:45.400985 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 0.684667 (* 0.0272727 = 0.0186727 loss) | |
| I0506 00:39:45.401000 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.141235 (* 0.0272727 = 0.00385185 loss) | |
| I0506 00:39:45.401012 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0736969 (* 0.0272727 = 0.00200991 loss) | |
| I0506 00:39:45.401026 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0569693 (* 0.0272727 = 0.00155371 loss) | |
| I0506 00:39:45.401039 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0519677 (* 0.0272727 = 0.0014173 loss) | |
| I0506 00:39:45.401052 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0398315 (* 0.0272727 = 0.00108631 loss) | |
| I0506 00:39:45.401065 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0360374 (* 0.0272727 = 0.000982839 loss) | |
| I0506 00:39:45.401079 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.0306675 (* 0.0272727 = 0.000836386 loss) | |
| I0506 00:39:45.401093 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.0208196 (* 0.0272727 = 0.000567807 loss) | |
| I0506 00:39:45.401105 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0157995 (* 0.0272727 = 0.000430894 loss) | |
| I0506 00:39:45.401135 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.014277 (* 0.0272727 = 0.000389374 loss) | |
| I0506 00:39:45.401152 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0137202 (* 0.0272727 = 0.000374189 loss) | |
| I0506 00:39:45.401166 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0132576 (* 0.0272727 = 0.000361572 loss) | |
| I0506 00:39:45.401180 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0128908 (* 0.0272727 = 0.000351567 loss) | |
| I0506 00:39:45.401193 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0130262 (* 0.0272727 = 0.000355259 loss) | |
| I0506 00:39:45.401206 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0555345 | |
| I0506 00:39:45.401217 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.09 | |
| I0506 00:39:45.401228 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.083 | |
| I0506 00:39:45.401239 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.055 | |
| I0506 00:39:45.401250 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.143 | |
| I0506 00:39:45.401262 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.284 | |
| I0506 00:39:45.401273 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.446 | |
| I0506 00:39:45.401284 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.729 | |
| I0506 00:39:45.401295 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.91 | |
| I0506 00:39:45.401306 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.99 | |
| I0506 00:39:45.401317 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.999 | |
| I0506 00:39:45.401329 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0506 00:39:45.401340 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0506 00:39:45.401350 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0506 00:39:45.401361 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0506 00:39:45.401371 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0506 00:39:45.401382 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0506 00:39:45.401393 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0506 00:39:45.401417 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0506 00:39:45.401428 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0506 00:39:45.401439 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0506 00:39:45.401454 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 00:39:45.401475 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 00:39:45.401494 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.758865 | |
| I0506 00:39:45.401506 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.199386 | |
| I0506 00:39:45.401520 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.54302 (* 1 = 3.54302 loss) | |
| I0506 00:39:45.401533 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.965883 (* 1 = 0.965883 loss) | |
| I0506 00:39:45.401547 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 3.45983 (* 0.0909091 = 0.31453 loss) | |
| I0506 00:39:45.401561 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 3.52331 (* 0.0909091 = 0.320301 loss) | |
| I0506 00:39:45.401573 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.60732 (* 0.0909091 = 0.327938 loss) | |
| I0506 00:39:45.401587 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 3.41953 (* 0.0909091 = 0.310866 loss) | |
| I0506 00:39:45.401599 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 3.08296 (* 0.0909091 = 0.280269 loss) | |
| I0506 00:39:45.401612 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.62448 (* 0.0909091 = 0.238589 loss) | |
| I0506 00:39:45.401625 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.54298 (* 0.0909091 = 0.140271 loss) | |
| I0506 00:39:45.401638 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.600824 (* 0.0909091 = 0.0546204 loss) | |
| I0506 00:39:45.401653 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.0874537 (* 0.0909091 = 0.00795034 loss) | |
| I0506 00:39:45.401665 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0445834 (* 0.0909091 = 0.00405304 loss) | |
| I0506 00:39:45.401679 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0333765 (* 0.0909091 = 0.00303423 loss) | |
| I0506 00:39:45.401692 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.027383 (* 0.0909091 = 0.00248936 loss) | |
| I0506 00:39:45.401705 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0236712 (* 0.0909091 = 0.00215193 loss) | |
| I0506 00:39:45.401718 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.019458 (* 0.0909091 = 0.00176891 loss) | |
| I0506 00:39:45.401736 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.0144217 (* 0.0909091 = 0.00131107 loss) | |
| I0506 00:39:45.401751 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.0085763 (* 0.0909091 = 0.000779664 loss) | |
| I0506 00:39:45.401764 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00415313 (* 0.0909091 = 0.000377558 loss) | |
| I0506 00:39:45.401777 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00327481 (* 0.0909091 = 0.00029771 loss) | |
| I0506 00:39:45.401792 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00238702 (* 0.0909091 = 0.000217002 loss) | |
| I0506 00:39:45.401804 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00203339 (* 0.0909091 = 0.000184854 loss) | |
| I0506 00:39:45.401818 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0017253 (* 0.0909091 = 0.000156846 loss) | |
| I0506 00:39:45.401831 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00137777 (* 0.0909091 = 0.000125252 loss) | |
| I0506 00:39:45.401844 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 00:39:45.401854 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:39:45.401865 15760 solver.cpp:406] Test net output #149: total_confidence = 0.000108748 | |
| I0506 00:39:45.401876 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000126195 | |
| I0506 00:39:45.401895 15760 solver.cpp:338] Iteration 5000, Testing net (#1) | |
| I0506 00:40:22.368854 15760 solver.cpp:393] Test loss: 11.9666 | |
| I0506 00:40:22.368981 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0288351 | |
| I0506 00:40:22.369000 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.065 | |
| I0506 00:40:22.369014 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.06 | |
| I0506 00:40:22.369026 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.034 | |
| I0506 00:40:22.369038 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.124 | |
| I0506 00:40:22.369050 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.25 | |
| I0506 00:40:22.369061 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.407 | |
| I0506 00:40:22.369073 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.633 | |
| I0506 00:40:22.369093 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.792 | |
| I0506 00:40:22.369104 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.884 | |
| I0506 00:40:22.369117 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.909 | |
| I0506 00:40:22.369145 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.924 | |
| I0506 00:40:22.369166 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.935 | |
| I0506 00:40:22.369177 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.948 | |
| I0506 00:40:22.369189 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.959 | |
| I0506 00:40:22.369200 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.969 | |
| I0506 00:40:22.369212 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.978 | |
| I0506 00:40:22.369225 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.989 | |
| I0506 00:40:22.369235 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0506 00:40:22.369247 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0506 00:40:22.369259 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.999 | |
| I0506 00:40:22.369271 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 00:40:22.369282 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 00:40:22.369293 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.720182 | |
| I0506 00:40:22.369305 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.129125 | |
| I0506 00:40:22.369323 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 4.55856 (* 0.3 = 1.36757 loss) | |
| I0506 00:40:22.369336 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.38516 (* 0.3 = 0.415549 loss) | |
| I0506 00:40:22.369350 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.66958 (* 0.0272727 = 0.10008 loss) | |
| I0506 00:40:22.369367 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.68548 (* 0.0272727 = 0.100513 loss) | |
| I0506 00:40:22.369380 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.76582 (* 0.0272727 = 0.102704 loss) | |
| I0506 00:40:22.369393 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.63249 (* 0.0272727 = 0.0990679 loss) | |
| I0506 00:40:22.369406 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 3.2862 (* 0.0272727 = 0.0896236 loss) | |
| I0506 00:40:22.369428 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.83911 (* 0.0272727 = 0.0774302 loss) | |
| I0506 00:40:22.369441 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.96588 (* 0.0272727 = 0.0536148 loss) | |
| I0506 00:40:22.369456 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 1.19099 (* 0.0272727 = 0.0324815 loss) | |
| I0506 00:40:22.369468 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.681106 (* 0.0272727 = 0.0185756 loss) | |
| I0506 00:40:22.369482 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.577888 (* 0.0272727 = 0.0157606 loss) | |
| I0506 00:40:22.369494 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.494281 (* 0.0272727 = 0.0134804 loss) | |
| I0506 00:40:22.369508 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.443227 (* 0.0272727 = 0.012088 loss) | |
| I0506 00:40:22.369521 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.370128 (* 0.0272727 = 0.0100944 loss) | |
| I0506 00:40:22.369554 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.307017 (* 0.0272727 = 0.0083732 loss) | |
| I0506 00:40:22.369570 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.238482 (* 0.0272727 = 0.00650405 loss) | |
| I0506 00:40:22.369583 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.174826 (* 0.0272727 = 0.00476798 loss) | |
| I0506 00:40:22.369596 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0996975 (* 0.0272727 = 0.00271902 loss) | |
| I0506 00:40:22.369609 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0490802 (* 0.0272727 = 0.00133855 loss) | |
| I0506 00:40:22.369623 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0409047 (* 0.0272727 = 0.00111558 loss) | |
| I0506 00:40:22.369637 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0269477 (* 0.0272727 = 0.000734938 loss) | |
| I0506 00:40:22.369650 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0172787 (* 0.0272727 = 0.000471238 loss) | |
| I0506 00:40:22.369663 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0171458 (* 0.0272727 = 0.000467612 loss) | |
| I0506 00:40:22.369675 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0232811 | |
| I0506 00:40:22.369686 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.065 | |
| I0506 00:40:22.369699 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.064 | |
| I0506 00:40:22.369709 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.039 | |
| I0506 00:40:22.369724 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.117 | |
| I0506 00:40:22.369737 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.25 | |
| I0506 00:40:22.369748 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.407 | |
| I0506 00:40:22.369760 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.633 | |
| I0506 00:40:22.369771 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.792 | |
| I0506 00:40:22.369781 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.884 | |
| I0506 00:40:22.369792 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.909 | |
| I0506 00:40:22.369804 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.924 | |
| I0506 00:40:22.369815 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.935 | |
| I0506 00:40:22.369827 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.948 | |
| I0506 00:40:22.369837 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.959 | |
| I0506 00:40:22.369848 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.969 | |
| I0506 00:40:22.369859 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.978 | |
| I0506 00:40:22.369871 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.989 | |
| I0506 00:40:22.369882 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0506 00:40:22.369894 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0506 00:40:22.369904 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.999 | |
| I0506 00:40:22.369915 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 00:40:22.369930 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 00:40:22.369941 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.719136 | |
| I0506 00:40:22.369953 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.134291 | |
| I0506 00:40:22.369963 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 4.52435 (* 0.3 = 1.3573 loss) | |
| I0506 00:40:22.369973 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.37004 (* 0.3 = 0.411011 loss) | |
| I0506 00:40:22.369987 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 3.66483 (* 0.0272727 = 0.0999499 loss) | |
| I0506 00:40:22.370002 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.67993 (* 0.0272727 = 0.100362 loss) | |
| I0506 00:40:22.370025 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.7498 (* 0.0272727 = 0.102267 loss) | |
| I0506 00:40:22.370044 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.63579 (* 0.0272727 = 0.0991579 loss) | |
| I0506 00:40:22.370057 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 3.29958 (* 0.0272727 = 0.0899887 loss) | |
| I0506 00:40:22.370070 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.82668 (* 0.0272727 = 0.0770912 loss) | |
| I0506 00:40:22.370084 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.97877 (* 0.0272727 = 0.0539664 loss) | |
| I0506 00:40:22.370096 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 1.19632 (* 0.0272727 = 0.0326269 loss) | |
| I0506 00:40:22.370110 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.689656 (* 0.0272727 = 0.0188088 loss) | |
| I0506 00:40:22.370122 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.565355 (* 0.0272727 = 0.0154188 loss) | |
| I0506 00:40:22.370136 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.49978 (* 0.0272727 = 0.0136304 loss) | |
| I0506 00:40:22.370148 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.442456 (* 0.0272727 = 0.012067 loss) | |
| I0506 00:40:22.370162 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.367251 (* 0.0272727 = 0.0100159 loss) | |
| I0506 00:40:22.370174 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.301094 (* 0.0272727 = 0.00821165 loss) | |
| I0506 00:40:22.370187 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.233142 (* 0.0272727 = 0.00635841 loss) | |
| I0506 00:40:22.370200 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.17466 (* 0.0272727 = 0.00476346 loss) | |
| I0506 00:40:22.370213 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0960569 (* 0.0272727 = 0.00261973 loss) | |
| I0506 00:40:22.370226 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0445587 (* 0.0272727 = 0.00121524 loss) | |
| I0506 00:40:22.370239 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0368121 (* 0.0272727 = 0.00100397 loss) | |
| I0506 00:40:22.370252 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0223634 (* 0.0272727 = 0.000609911 loss) | |
| I0506 00:40:22.370265 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0139406 (* 0.0272727 = 0.000380199 loss) | |
| I0506 00:40:22.370278 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0134886 (* 0.0272727 = 0.000367872 loss) | |
| I0506 00:40:22.370290 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0544308 | |
| I0506 00:40:22.370301 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.09 | |
| I0506 00:40:22.370312 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.083 | |
| I0506 00:40:22.370323 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.062 | |
| I0506 00:40:22.370334 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.147 | |
| I0506 00:40:22.370345 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.267 | |
| I0506 00:40:22.370357 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.407 | |
| I0506 00:40:22.370368 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.633 | |
| I0506 00:40:22.370378 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.792 | |
| I0506 00:40:22.370389 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.884 | |
| I0506 00:40:22.370400 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.909 | |
| I0506 00:40:22.370411 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.924 | |
| I0506 00:40:22.370422 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.935 | |
| I0506 00:40:22.370432 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.948 | |
| I0506 00:40:22.370443 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.959 | |
| I0506 00:40:22.370455 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.969 | |
| I0506 00:40:22.370466 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.978 | |
| I0506 00:40:22.370486 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.989 | |
| I0506 00:40:22.370498 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0506 00:40:22.370509 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0506 00:40:22.370520 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.999 | |
| I0506 00:40:22.370532 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 00:40:22.370543 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 00:40:22.370554 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.725 | |
| I0506 00:40:22.370565 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.191833 | |
| I0506 00:40:22.370579 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.51349 (* 1 = 3.51349 loss) | |
| I0506 00:40:22.370591 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.09993 (* 1 = 1.09993 loss) | |
| I0506 00:40:22.370604 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 3.38436 (* 0.0909091 = 0.307669 loss) | |
| I0506 00:40:22.370617 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 3.45736 (* 0.0909091 = 0.314305 loss) | |
| I0506 00:40:22.370630 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.54114 (* 0.0909091 = 0.321922 loss) | |
| I0506 00:40:22.370643 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 3.39087 (* 0.0909091 = 0.308261 loss) | |
| I0506 00:40:22.370656 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 3.04134 (* 0.0909091 = 0.276486 loss) | |
| I0506 00:40:22.370669 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.63886 (* 0.0909091 = 0.239896 loss) | |
| I0506 00:40:22.370682 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.82592 (* 0.0909091 = 0.165993 loss) | |
| I0506 00:40:22.370695 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 1.06301 (* 0.0909091 = 0.096637 loss) | |
| I0506 00:40:22.370708 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.563858 (* 0.0909091 = 0.0512598 loss) | |
| I0506 00:40:22.370721 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.463761 (* 0.0909091 = 0.0421601 loss) | |
| I0506 00:40:22.370734 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.412728 (* 0.0909091 = 0.0375207 loss) | |
| I0506 00:40:22.370748 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.377764 (* 0.0909091 = 0.0343421 loss) | |
| I0506 00:40:22.370760 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.315398 (* 0.0909091 = 0.0286725 loss) | |
| I0506 00:40:22.370779 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.260061 (* 0.0909091 = 0.0236419 loss) | |
| I0506 00:40:22.370792 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.206498 (* 0.0909091 = 0.0187726 loss) | |
| I0506 00:40:22.370805 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.164701 (* 0.0909091 = 0.0149729 loss) | |
| I0506 00:40:22.370820 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0939701 (* 0.0909091 = 0.00854273 loss) | |
| I0506 00:40:22.370832 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0404725 (* 0.0909091 = 0.00367931 loss) | |
| I0506 00:40:22.370846 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0302107 (* 0.0909091 = 0.00274643 loss) | |
| I0506 00:40:22.370859 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0116014 (* 0.0909091 = 0.00105467 loss) | |
| I0506 00:40:22.370872 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00179832 (* 0.0909091 = 0.000163484 loss) | |
| I0506 00:40:22.370887 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00141692 (* 0.0909091 = 0.000128811 loss) | |
| I0506 00:40:22.370898 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 00:40:22.370908 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:40:22.370919 15760 solver.cpp:406] Test net output #149: total_confidence = 9.15408e-05 | |
| I0506 00:40:22.370940 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 9.53168e-05 | |
| I0506 00:40:22.508584 15760 solver.cpp:229] Iteration 5000, loss = 11.6962 | |
| I0506 00:40:22.508651 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:40:22.508669 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:40:22.508682 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 00:40:22.508694 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 00:40:22.508707 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:40:22.508718 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:40:22.508729 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 00:40:22.508741 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:40:22.508754 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:40:22.508765 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:40:22.508776 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:40:22.508788 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:40:22.508800 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:40:22.508810 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:40:22.508822 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:40:22.508834 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:40:22.508846 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:40:22.508860 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:40:22.508872 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:40:22.508883 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:40:22.508894 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:40:22.508906 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:40:22.508918 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:40:22.508929 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0506 00:40:22.508940 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.06 | |
| I0506 00:40:22.508956 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.74631 (* 0.3 = 1.12389 loss) | |
| I0506 00:40:22.508971 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.15464 (* 0.3 = 0.346392 loss) | |
| I0506 00:40:22.508985 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.69347 (* 0.0272727 = 0.100731 loss) | |
| I0506 00:40:22.508998 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.77612 (* 0.0272727 = 0.102985 loss) | |
| I0506 00:40:22.509012 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.29957 (* 0.0272727 = 0.0899884 loss) | |
| I0506 00:40:22.509026 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.83759 (* 0.0272727 = 0.104661 loss) | |
| I0506 00:40:22.509039 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.52195 (* 0.0272727 = 0.0960531 loss) | |
| I0506 00:40:22.509052 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.64183 (* 0.0272727 = 0.0993226 loss) | |
| I0506 00:40:22.509066 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 3.2276 (* 0.0272727 = 0.0880255 loss) | |
| I0506 00:40:22.509079 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.8893 (* 0.0272727 = 0.0242536 loss) | |
| I0506 00:40:22.509093 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0266088 (* 0.0272727 = 0.000725696 loss) | |
| I0506 00:40:22.509116 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0282175 (* 0.0272727 = 0.000769569 loss) | |
| I0506 00:40:22.509147 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0315762 (* 0.0272727 = 0.000861169 loss) | |
| I0506 00:40:22.509201 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0242613 (* 0.0272727 = 0.000661672 loss) | |
| I0506 00:40:22.509215 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0227553 (* 0.0272727 = 0.000620599 loss) | |
| I0506 00:40:22.509229 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0172005 (* 0.0272727 = 0.000469105 loss) | |
| I0506 00:40:22.509243 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.019726 (* 0.0272727 = 0.000537981 loss) | |
| I0506 00:40:22.509256 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00898979 (* 0.0272727 = 0.000245176 loss) | |
| I0506 00:40:22.509270 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00410569 (* 0.0272727 = 0.000111973 loss) | |
| I0506 00:40:22.509284 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00686188 (* 0.0272727 = 0.000187142 loss) | |
| I0506 00:40:22.509297 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0042966 (* 0.0272727 = 0.00011718 loss) | |
| I0506 00:40:22.509310 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00462673 (* 0.0272727 = 0.000126183 loss) | |
| I0506 00:40:22.509325 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00419469 (* 0.0272727 = 0.000114401 loss) | |
| I0506 00:40:22.509337 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00847785 (* 0.0272727 = 0.000231214 loss) | |
| I0506 00:40:22.509349 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.02 | |
| I0506 00:40:22.509361 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:40:22.509369 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:40:22.509377 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 00:40:22.509384 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 00:40:22.509397 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 00:40:22.509407 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 00:40:22.509419 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:40:22.509431 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:40:22.509443 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:40:22.509454 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:40:22.509465 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:40:22.509476 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:40:22.509487 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:40:22.509498 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:40:22.509510 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:40:22.509521 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:40:22.509531 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:40:22.509542 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:40:22.509553 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:40:22.509564 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:40:22.509575 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:40:22.509587 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:40:22.509598 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.721591 | |
| I0506 00:40:22.509609 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.1 | |
| I0506 00:40:22.509623 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.79315 (* 0.3 = 1.13795 loss) | |
| I0506 00:40:22.509636 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.1525 (* 0.3 = 0.345751 loss) | |
| I0506 00:40:22.509650 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.08041 (* 0.0272727 = 0.111284 loss) | |
| I0506 00:40:22.509682 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.09387 (* 0.0272727 = 0.111651 loss) | |
| I0506 00:40:22.509799 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.52455 (* 0.0272727 = 0.096124 loss) | |
| I0506 00:40:22.509820 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.08873 (* 0.0272727 = 0.111511 loss) | |
| I0506 00:40:22.509834 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.85967 (* 0.0272727 = 0.105264 loss) | |
| I0506 00:40:22.509847 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.12843 (* 0.0272727 = 0.0853207 loss) | |
| I0506 00:40:22.509861 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 3.00106 (* 0.0272727 = 0.0818472 loss) | |
| I0506 00:40:22.509882 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.16671 (* 0.0272727 = 0.0318194 loss) | |
| I0506 00:40:22.509908 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0723691 (* 0.0272727 = 0.0019737 loss) | |
| I0506 00:40:22.509953 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0496581 (* 0.0272727 = 0.00135431 loss) | |
| I0506 00:40:22.509970 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0382053 (* 0.0272727 = 0.00104196 loss) | |
| I0506 00:40:22.509985 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0352455 (* 0.0272727 = 0.00096124 loss) | |
| I0506 00:40:22.509999 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0221103 (* 0.0272727 = 0.000603007 loss) | |
| I0506 00:40:22.510020 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0185505 (* 0.0272727 = 0.000505924 loss) | |
| I0506 00:40:22.510051 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0175232 (* 0.0272727 = 0.000477906 loss) | |
| I0506 00:40:22.510066 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0126316 (* 0.0272727 = 0.000344499 loss) | |
| I0506 00:40:22.510088 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0123698 (* 0.0272727 = 0.000337358 loss) | |
| I0506 00:40:22.510102 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00707177 (* 0.0272727 = 0.000192866 loss) | |
| I0506 00:40:22.510116 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00937575 (* 0.0272727 = 0.000255702 loss) | |
| I0506 00:40:22.510129 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00899948 (* 0.0272727 = 0.00024544 loss) | |
| I0506 00:40:22.510151 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00827024 (* 0.0272727 = 0.000225552 loss) | |
| I0506 00:40:22.510164 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00614708 (* 0.0272727 = 0.000167648 loss) | |
| I0506 00:40:22.510176 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0506 00:40:22.510187 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:40:22.510208 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:40:22.510220 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:40:22.510231 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:40:22.510241 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:40:22.510253 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 00:40:22.510288 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:40:22.510301 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:40:22.510313 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:40:22.510332 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:40:22.510344 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:40:22.510355 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:40:22.510366 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:40:22.510377 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:40:22.510401 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:40:22.510413 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:40:22.510426 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:40:22.510437 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:40:22.510447 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:40:22.510458 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:40:22.510469 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:40:22.510480 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:40:22.510491 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.715909 | |
| I0506 00:40:22.510504 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.14 | |
| I0506 00:40:22.510516 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.58243 (* 1 = 3.58243 loss) | |
| I0506 00:40:22.510530 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.09348 (* 1 = 1.09348 loss) | |
| I0506 00:40:22.510545 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.82456 (* 0.0909091 = 0.347688 loss) | |
| I0506 00:40:22.510557 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.62655 (* 0.0909091 = 0.329686 loss) | |
| I0506 00:40:22.510571 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.43081 (* 0.0909091 = 0.311892 loss) | |
| I0506 00:40:22.510584 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.38981 (* 0.0909091 = 0.308165 loss) | |
| I0506 00:40:22.510597 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.20746 (* 0.0909091 = 0.291587 loss) | |
| I0506 00:40:22.510612 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.35644 (* 0.0909091 = 0.305131 loss) | |
| I0506 00:40:22.510624 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 3.11747 (* 0.0909091 = 0.283406 loss) | |
| I0506 00:40:22.510637 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.848095 (* 0.0909091 = 0.0770996 loss) | |
| I0506 00:40:22.510651 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.013667 (* 0.0909091 = 0.00124245 loss) | |
| I0506 00:40:22.510665 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0177721 (* 0.0909091 = 0.00161564 loss) | |
| I0506 00:40:22.510679 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0125938 (* 0.0909091 = 0.00114489 loss) | |
| I0506 00:40:22.510692 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0117343 (* 0.0909091 = 0.00106676 loss) | |
| I0506 00:40:22.510705 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00927592 (* 0.0909091 = 0.000843265 loss) | |
| I0506 00:40:22.510718 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00850808 (* 0.0909091 = 0.000773462 loss) | |
| I0506 00:40:22.510732 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00558495 (* 0.0909091 = 0.000507723 loss) | |
| I0506 00:40:22.510746 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00375123 (* 0.0909091 = 0.000341021 loss) | |
| I0506 00:40:22.510764 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00158938 (* 0.0909091 = 0.000144489 loss) | |
| I0506 00:40:22.510779 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00106593 (* 0.0909091 = 9.69025e-05 loss) | |
| I0506 00:40:22.510793 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000697817 (* 0.0909091 = 6.34379e-05 loss) | |
| I0506 00:40:22.510807 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000764633 (* 0.0909091 = 6.95121e-05 loss) | |
| I0506 00:40:22.510821 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000507627 (* 0.0909091 = 4.61479e-05 loss) | |
| I0506 00:40:22.510864 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000314227 (* 0.0909091 = 2.85661e-05 loss) | |
| I0506 00:40:22.510879 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:40:22.510910 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:40:22.510921 15760 solver.cpp:245] Train net output #149: total_confidence = 1.94226e-05 | |
| I0506 00:40:22.510933 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 6.34781e-05 | |
| I0506 00:40:22.510946 15760 sgd_solver.cpp:106] Iteration 5000, lr = 0.001 | |
| I0506 00:40:55.341235 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.6563 > 30) by scale factor 0.841367 | |
| I0506 00:41:28.324769 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.6717 > 30) by scale factor 0.841003 | |
| I0506 00:41:30.056152 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.9235 > 30) by scale factor 0.939747 | |
| I0506 00:42:10.325240 15760 solver.cpp:229] Iteration 5500, loss = 11.6692 | |
| I0506 00:42:10.325407 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0185185 | |
| I0506 00:42:10.325428 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:42:10.325441 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:42:10.325461 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:42:10.325474 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 00:42:10.325485 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:42:10.325497 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 00:42:10.325508 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:42:10.325521 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 00:42:10.325533 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 00:42:10.325546 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:42:10.325556 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:42:10.325568 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:42:10.325579 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:42:10.325592 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:42:10.325618 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:42:10.325630 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:42:10.325641 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:42:10.325652 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:42:10.325664 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:42:10.325680 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:42:10.325690 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:42:10.325701 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:42:10.325713 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.698864 | |
| I0506 00:42:10.325724 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.037037 | |
| I0506 00:42:10.325742 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.77336 (* 0.3 = 1.13201 loss) | |
| I0506 00:42:10.325757 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.3148 (* 0.3 = 0.39444 loss) | |
| I0506 00:42:10.325770 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.56202 (* 0.0272727 = 0.0971459 loss) | |
| I0506 00:42:10.325784 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.39636 (* 0.0272727 = 0.0926281 loss) | |
| I0506 00:42:10.325798 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.30526 (* 0.0272727 = 0.117416 loss) | |
| I0506 00:42:10.325811 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.88276 (* 0.0272727 = 0.105894 loss) | |
| I0506 00:42:10.325825 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.60972 (* 0.0272727 = 0.0984469 loss) | |
| I0506 00:42:10.325839 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.4883 (* 0.0272727 = 0.0678628 loss) | |
| I0506 00:42:10.325851 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.47182 (* 0.0272727 = 0.0674131 loss) | |
| I0506 00:42:10.325865 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 2.0965 (* 0.0272727 = 0.0571773 loss) | |
| I0506 00:42:10.325882 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.23953 (* 0.0272727 = 0.0338054 loss) | |
| I0506 00:42:10.325897 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.686877 (* 0.0272727 = 0.018733 loss) | |
| I0506 00:42:10.325911 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.872287 (* 0.0272727 = 0.0237896 loss) | |
| I0506 00:42:10.325925 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0845576 (* 0.0272727 = 0.00230612 loss) | |
| I0506 00:42:10.325939 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0786655 (* 0.0272727 = 0.00214542 loss) | |
| I0506 00:42:10.325979 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0570378 (* 0.0272727 = 0.00155558 loss) | |
| I0506 00:42:10.325995 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0439905 (* 0.0272727 = 0.00119974 loss) | |
| I0506 00:42:10.326009 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0248823 (* 0.0272727 = 0.000678609 loss) | |
| I0506 00:42:10.326027 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0120519 (* 0.0272727 = 0.000328687 loss) | |
| I0506 00:42:10.326042 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0114849 (* 0.0272727 = 0.000313225 loss) | |
| I0506 00:42:10.326056 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0184248 (* 0.0272727 = 0.000502494 loss) | |
| I0506 00:42:10.326078 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0263443 (* 0.0272727 = 0.000718481 loss) | |
| I0506 00:42:10.326092 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0112407 (* 0.0272727 = 0.000306566 loss) | |
| I0506 00:42:10.326105 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0216104 (* 0.0272727 = 0.000589375 loss) | |
| I0506 00:42:10.326117 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:42:10.326129 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:42:10.326140 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 00:42:10.326153 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:42:10.326164 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:42:10.326174 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 00:42:10.326190 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 00:42:10.326202 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:42:10.326213 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0506 00:42:10.326225 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 00:42:10.326236 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:42:10.326247 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:42:10.326264 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:42:10.326275 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:42:10.326287 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:42:10.326297 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:42:10.326308 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:42:10.326319 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:42:10.326330 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:42:10.326341 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:42:10.326354 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:42:10.326364 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:42:10.326375 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:42:10.326386 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0506 00:42:10.326401 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.037037 | |
| I0506 00:42:10.326421 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.81862 (* 0.3 = 1.14559 loss) | |
| I0506 00:42:10.326436 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.38362 (* 0.3 = 0.415085 loss) | |
| I0506 00:42:10.326448 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.92069 (* 0.0272727 = 0.106928 loss) | |
| I0506 00:42:10.326467 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.43389 (* 0.0272727 = 0.0936516 loss) | |
| I0506 00:42:10.326493 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.86065 (* 0.0272727 = 0.10529 loss) | |
| I0506 00:42:10.326506 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.30845 (* 0.0272727 = 0.0902304 loss) | |
| I0506 00:42:10.326520 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.2967 (* 0.0272727 = 0.0899099 loss) | |
| I0506 00:42:10.326534 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.86538 (* 0.0272727 = 0.0781469 loss) | |
| I0506 00:42:10.326547 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.40633 (* 0.0272727 = 0.0656271 loss) | |
| I0506 00:42:10.326561 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.83471 (* 0.0272727 = 0.0500374 loss) | |
| I0506 00:42:10.326575 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.22599 (* 0.0272727 = 0.0334361 loss) | |
| I0506 00:42:10.326588 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.670039 (* 0.0272727 = 0.0182738 loss) | |
| I0506 00:42:10.326601 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.745505 (* 0.0272727 = 0.0203319 loss) | |
| I0506 00:42:10.326616 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0925337 (* 0.0272727 = 0.00252365 loss) | |
| I0506 00:42:10.326629 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0410184 (* 0.0272727 = 0.00111868 loss) | |
| I0506 00:42:10.326642 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.04672 (* 0.0272727 = 0.00127418 loss) | |
| I0506 00:42:10.326656 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.025204 (* 0.0272727 = 0.000687383 loss) | |
| I0506 00:42:10.326670 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0266322 (* 0.0272727 = 0.000726333 loss) | |
| I0506 00:42:10.326684 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0221192 (* 0.0272727 = 0.000603251 loss) | |
| I0506 00:42:10.326697 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0194918 (* 0.0272727 = 0.000531594 loss) | |
| I0506 00:42:10.326710 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0173857 (* 0.0272727 = 0.000474155 loss) | |
| I0506 00:42:10.326725 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0107142 (* 0.0272727 = 0.000292206 loss) | |
| I0506 00:42:10.326738 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0120979 (* 0.0272727 = 0.000329944 loss) | |
| I0506 00:42:10.326751 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0133348 (* 0.0272727 = 0.000363677 loss) | |
| I0506 00:42:10.326763 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.037037 | |
| I0506 00:42:10.326776 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 00:42:10.326786 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 00:42:10.326798 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 00:42:10.326809 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:42:10.326820 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:42:10.326833 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 00:42:10.326843 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:42:10.326854 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0506 00:42:10.326865 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 00:42:10.326877 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:42:10.326889 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:42:10.326900 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:42:10.326911 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:42:10.326922 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:42:10.326941 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:42:10.326952 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:42:10.326974 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:42:10.326987 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:42:10.326998 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:42:10.327008 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:42:10.327034 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:42:10.327046 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:42:10.327061 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0506 00:42:10.327074 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.203704 | |
| I0506 00:42:10.327088 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.60833 (* 1 = 3.60833 loss) | |
| I0506 00:42:10.327102 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.22241 (* 1 = 1.22241 loss) | |
| I0506 00:42:10.327116 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.60434 (* 0.0909091 = 0.327668 loss) | |
| I0506 00:42:10.327128 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.57342 (* 0.0909091 = 0.324856 loss) | |
| I0506 00:42:10.327142 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.50442 (* 0.0909091 = 0.318584 loss) | |
| I0506 00:42:10.327162 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.09721 (* 0.0909091 = 0.281565 loss) | |
| I0506 00:42:10.327174 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.51152 (* 0.0909091 = 0.319229 loss) | |
| I0506 00:42:10.327188 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.73252 (* 0.0909091 = 0.248411 loss) | |
| I0506 00:42:10.327201 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.34001 (* 0.0909091 = 0.212728 loss) | |
| I0506 00:42:10.327214 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.70367 (* 0.0909091 = 0.154879 loss) | |
| I0506 00:42:10.327234 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.01027 (* 0.0909091 = 0.0918427 loss) | |
| I0506 00:42:10.327249 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.723234 (* 0.0909091 = 0.0657485 loss) | |
| I0506 00:42:10.327261 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.692879 (* 0.0909091 = 0.062989 loss) | |
| I0506 00:42:10.327275 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0835956 (* 0.0909091 = 0.0075996 loss) | |
| I0506 00:42:10.327288 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0433524 (* 0.0909091 = 0.00394112 loss) | |
| I0506 00:42:10.327301 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0347878 (* 0.0909091 = 0.00316253 loss) | |
| I0506 00:42:10.327314 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0245039 (* 0.0909091 = 0.00222763 loss) | |
| I0506 00:42:10.327328 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0188591 (* 0.0909091 = 0.00171446 loss) | |
| I0506 00:42:10.327342 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0060941 (* 0.0909091 = 0.000554009 loss) | |
| I0506 00:42:10.327355 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00378667 (* 0.0909091 = 0.000344243 loss) | |
| I0506 00:42:10.327368 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00265264 (* 0.0909091 = 0.000241149 loss) | |
| I0506 00:42:10.327383 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00186553 (* 0.0909091 = 0.000169593 loss) | |
| I0506 00:42:10.327396 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00106786 (* 0.0909091 = 9.70784e-05 loss) | |
| I0506 00:42:10.327409 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00141573 (* 0.0909091 = 0.000128703 loss) | |
| I0506 00:42:10.327421 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:42:10.327432 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:42:10.327445 15760 solver.cpp:245] Train net output #149: total_confidence = 1.23584e-05 | |
| I0506 00:42:10.327471 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000267152 | |
| I0506 00:42:10.327486 15760 sgd_solver.cpp:106] Iteration 5500, lr = 0.001 | |
| I0506 00:43:11.496431 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.0541 > 30) by scale factor 0.966056 | |
| I0506 00:43:57.778695 15760 solver.cpp:229] Iteration 6000, loss = 11.4742 | |
| I0506 00:43:57.778812 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0810811 | |
| I0506 00:43:57.778831 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:43:57.778847 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:43:57.778861 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:43:57.778872 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0506 00:43:57.778883 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0506 00:43:57.778894 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 00:43:57.778906 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 00:43:57.778918 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 00:43:57.778929 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:43:57.778941 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:43:57.778955 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:43:57.778967 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:43:57.778980 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:43:57.778990 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:43:57.779002 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:43:57.779013 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:43:57.779026 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:43:57.779037 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:43:57.779048 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:43:57.779059 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:43:57.779072 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:43:57.779083 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:43:57.779093 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.778409 | |
| I0506 00:43:57.779105 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.162162 | |
| I0506 00:43:57.779121 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.58998 (* 0.3 = 1.07699 loss) | |
| I0506 00:43:57.779135 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.1054 (* 0.3 = 0.331621 loss) | |
| I0506 00:43:57.779150 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.57522 (* 0.0272727 = 0.097506 loss) | |
| I0506 00:43:57.779163 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.8534 (* 0.0272727 = 0.105093 loss) | |
| I0506 00:43:57.779177 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.50678 (* 0.0272727 = 0.0956395 loss) | |
| I0506 00:43:57.779191 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.14382 (* 0.0272727 = 0.0857404 loss) | |
| I0506 00:43:57.779204 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.71145 (* 0.0272727 = 0.0739487 loss) | |
| I0506 00:43:57.779217 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.96934 (* 0.0272727 = 0.0537092 loss) | |
| I0506 00:43:57.779232 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.11507 (* 0.0272727 = 0.030411 loss) | |
| I0506 00:43:57.779244 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.276354 (* 0.0272727 = 0.00753694 loss) | |
| I0506 00:43:57.779258 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.093735 (* 0.0272727 = 0.00255641 loss) | |
| I0506 00:43:57.779273 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0847181 (* 0.0272727 = 0.00231049 loss) | |
| I0506 00:43:57.779286 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0555037 (* 0.0272727 = 0.00151374 loss) | |
| I0506 00:43:57.779299 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0480605 (* 0.0272727 = 0.00131074 loss) | |
| I0506 00:43:57.779314 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0379006 (* 0.0272727 = 0.00103365 loss) | |
| I0506 00:43:57.779345 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0324523 (* 0.0272727 = 0.000885063 loss) | |
| I0506 00:43:57.779361 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0267234 (* 0.0272727 = 0.00072882 loss) | |
| I0506 00:43:57.779374 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0162839 (* 0.0272727 = 0.000444107 loss) | |
| I0506 00:43:57.779388 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0136602 (* 0.0272727 = 0.00037255 loss) | |
| I0506 00:43:57.779402 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0137741 (* 0.0272727 = 0.000375656 loss) | |
| I0506 00:43:57.779417 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00698953 (* 0.0272727 = 0.000190623 loss) | |
| I0506 00:43:57.779429 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00791223 (* 0.0272727 = 0.000215788 loss) | |
| I0506 00:43:57.779443 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00826333 (* 0.0272727 = 0.000225363 loss) | |
| I0506 00:43:57.779458 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00856035 (* 0.0272727 = 0.000233464 loss) | |
| I0506 00:43:57.779469 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0540541 | |
| I0506 00:43:57.779481 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0506 00:43:57.779494 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:43:57.779505 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 00:43:57.779515 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 00:43:57.779527 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 00:43:57.779538 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0506 00:43:57.779549 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 00:43:57.779561 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 00:43:57.779572 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:43:57.779583 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:43:57.779594 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:43:57.779606 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:43:57.779618 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:43:57.779629 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:43:57.779636 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:43:57.779644 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:43:57.779655 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:43:57.779666 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:43:57.779677 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:43:57.779688 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:43:57.779700 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:43:57.779711 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:43:57.779721 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0506 00:43:57.779732 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0810811 | |
| I0506 00:43:57.779747 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.73444 (* 0.3 = 1.12033 loss) | |
| I0506 00:43:57.779760 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.11469 (* 0.3 = 0.334408 loss) | |
| I0506 00:43:57.779774 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.21466 (* 0.0272727 = 0.0876726 loss) | |
| I0506 00:43:57.779788 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.79458 (* 0.0272727 = 0.103489 loss) | |
| I0506 00:43:57.779811 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.40568 (* 0.0272727 = 0.0928823 loss) | |
| I0506 00:43:57.779826 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.48387 (* 0.0272727 = 0.0950148 loss) | |
| I0506 00:43:57.779840 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.76113 (* 0.0272727 = 0.0753034 loss) | |
| I0506 00:43:57.779853 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.97485 (* 0.0272727 = 0.0538596 loss) | |
| I0506 00:43:57.779867 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.01428 (* 0.0272727 = 0.0276621 loss) | |
| I0506 00:43:57.779881 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.259536 (* 0.0272727 = 0.00707824 loss) | |
| I0506 00:43:57.779898 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.126123 (* 0.0272727 = 0.00343972 loss) | |
| I0506 00:43:57.779912 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0669654 (* 0.0272727 = 0.00182633 loss) | |
| I0506 00:43:57.779925 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0838407 (* 0.0272727 = 0.00228656 loss) | |
| I0506 00:43:57.779939 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0422779 (* 0.0272727 = 0.00115303 loss) | |
| I0506 00:43:57.779953 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0469876 (* 0.0272727 = 0.00128148 loss) | |
| I0506 00:43:57.779966 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0212913 (* 0.0272727 = 0.000580672 loss) | |
| I0506 00:43:57.779980 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.017784 (* 0.0272727 = 0.000485017 loss) | |
| I0506 00:43:57.779994 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0197501 (* 0.0272727 = 0.000538638 loss) | |
| I0506 00:43:57.780010 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00840394 (* 0.0272727 = 0.000229198 loss) | |
| I0506 00:43:57.780025 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00729914 (* 0.0272727 = 0.000199067 loss) | |
| I0506 00:43:57.780038 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00674832 (* 0.0272727 = 0.000184045 loss) | |
| I0506 00:43:57.780052 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00774124 (* 0.0272727 = 0.000211125 loss) | |
| I0506 00:43:57.780066 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.008188 (* 0.0272727 = 0.000223309 loss) | |
| I0506 00:43:57.780079 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00797718 (* 0.0272727 = 0.00021756 loss) | |
| I0506 00:43:57.780092 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.135135 | |
| I0506 00:43:57.780102 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 00:43:57.780114 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 00:43:57.780125 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:43:57.780136 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 00:43:57.780148 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 00:43:57.780160 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.75 | |
| I0506 00:43:57.780171 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0506 00:43:57.780182 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 00:43:57.780194 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:43:57.780205 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:43:57.780216 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:43:57.780227 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:43:57.780238 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:43:57.780249 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:43:57.780261 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:43:57.780272 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:43:57.780292 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:43:57.780305 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:43:57.780318 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:43:57.780328 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:43:57.780339 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:43:57.780350 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:43:57.780361 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.806818 | |
| I0506 00:43:57.780374 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.297297 | |
| I0506 00:43:57.780386 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.2338 (* 1 = 3.2338 loss) | |
| I0506 00:43:57.780400 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.839843 (* 1 = 0.839843 loss) | |
| I0506 00:43:57.780413 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.31735 (* 0.0909091 = 0.301577 loss) | |
| I0506 00:43:57.780427 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.74163 (* 0.0909091 = 0.340148 loss) | |
| I0506 00:43:57.780441 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.51769 (* 0.0909091 = 0.31979 loss) | |
| I0506 00:43:57.780453 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.48129 (* 0.0909091 = 0.316481 loss) | |
| I0506 00:43:57.780467 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.26178 (* 0.0909091 = 0.205616 loss) | |
| I0506 00:43:57.780480 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.46365 (* 0.0909091 = 0.133059 loss) | |
| I0506 00:43:57.780493 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.833514 (* 0.0909091 = 0.075774 loss) | |
| I0506 00:43:57.780506 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.187623 (* 0.0909091 = 0.0170566 loss) | |
| I0506 00:43:57.780520 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0427189 (* 0.0909091 = 0.00388354 loss) | |
| I0506 00:43:57.780534 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0395283 (* 0.0909091 = 0.00359348 loss) | |
| I0506 00:43:57.780547 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.02843 (* 0.0909091 = 0.00258454 loss) | |
| I0506 00:43:57.780560 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0341646 (* 0.0909091 = 0.00310587 loss) | |
| I0506 00:43:57.780575 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0278406 (* 0.0909091 = 0.00253097 loss) | |
| I0506 00:43:57.780587 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0309432 (* 0.0909091 = 0.00281302 loss) | |
| I0506 00:43:57.780601 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0182998 (* 0.0909091 = 0.00166362 loss) | |
| I0506 00:43:57.780614 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0153835 (* 0.0909091 = 0.0013985 loss) | |
| I0506 00:43:57.780628 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00769423 (* 0.0909091 = 0.000699475 loss) | |
| I0506 00:43:57.780642 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00646255 (* 0.0909091 = 0.000587504 loss) | |
| I0506 00:43:57.780655 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00301443 (* 0.0909091 = 0.000274039 loss) | |
| I0506 00:43:57.780668 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00282702 (* 0.0909091 = 0.000257002 loss) | |
| I0506 00:43:57.780683 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00210064 (* 0.0909091 = 0.000190967 loss) | |
| I0506 00:43:57.780696 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00228041 (* 0.0909091 = 0.00020731 loss) | |
| I0506 00:43:57.780707 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:43:57.780719 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:43:57.780730 15760 solver.cpp:245] Train net output #149: total_confidence = 9.13916e-06 | |
| I0506 00:43:57.780750 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000977473 | |
| I0506 00:43:57.780764 15760 sgd_solver.cpp:106] Iteration 6000, lr = 0.001 | |
| I0506 00:45:45.000713 15760 solver.cpp:229] Iteration 6500, loss = 11.1994 | |
| I0506 00:45:45.000830 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0465116 | |
| I0506 00:45:45.000850 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 00:45:45.000866 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:45:45.000880 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 00:45:45.000891 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 00:45:45.000903 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 00:45:45.000915 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 00:45:45.000926 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 00:45:45.000937 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:45:45.000949 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:45:45.000962 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:45:45.000972 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:45:45.000984 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:45:45.000995 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:45:45.001006 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:45:45.001018 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:45:45.001029 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:45:45.001041 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:45:45.001052 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:45:45.001071 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:45:45.001085 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:45:45.001108 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:45:45.001123 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:45:45.001135 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.75 | |
| I0506 00:45:45.001147 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.139535 | |
| I0506 00:45:45.001163 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.6682 (* 0.3 = 1.10046 loss) | |
| I0506 00:45:45.001178 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.1275 (* 0.3 = 0.338249 loss) | |
| I0506 00:45:45.001191 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.18083 (* 0.0272727 = 0.114023 loss) | |
| I0506 00:45:45.001205 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.74471 (* 0.0272727 = 0.102128 loss) | |
| I0506 00:45:45.001219 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.85873 (* 0.0272727 = 0.105238 loss) | |
| I0506 00:45:45.001232 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.26453 (* 0.0272727 = 0.0890325 loss) | |
| I0506 00:45:45.001245 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.79206 (* 0.0272727 = 0.0761472 loss) | |
| I0506 00:45:45.001258 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.13847 (* 0.0272727 = 0.0855945 loss) | |
| I0506 00:45:45.001272 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.32448 (* 0.0272727 = 0.0361222 loss) | |
| I0506 00:45:45.001286 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.841896 (* 0.0272727 = 0.0229608 loss) | |
| I0506 00:45:45.001299 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0763734 (* 0.0272727 = 0.00208291 loss) | |
| I0506 00:45:45.001313 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0634932 (* 0.0272727 = 0.00173163 loss) | |
| I0506 00:45:45.001327 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0335026 (* 0.0272727 = 0.000913707 loss) | |
| I0506 00:45:45.001341 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0475581 (* 0.0272727 = 0.00129704 loss) | |
| I0506 00:45:45.001354 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0164225 (* 0.0272727 = 0.000447887 loss) | |
| I0506 00:45:45.001386 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.02132 (* 0.0272727 = 0.000581456 loss) | |
| I0506 00:45:45.001402 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.019064 (* 0.0272727 = 0.000519927 loss) | |
| I0506 00:45:45.001415 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0210501 (* 0.0272727 = 0.000574093 loss) | |
| I0506 00:45:45.001430 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00979053 (* 0.0272727 = 0.000267014 loss) | |
| I0506 00:45:45.001444 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00880097 (* 0.0272727 = 0.000240026 loss) | |
| I0506 00:45:45.001457 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00541412 (* 0.0272727 = 0.000147658 loss) | |
| I0506 00:45:45.001471 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00782166 (* 0.0272727 = 0.000213318 loss) | |
| I0506 00:45:45.001485 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00853513 (* 0.0272727 = 0.000232776 loss) | |
| I0506 00:45:45.001505 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00675797 (* 0.0272727 = 0.000184308 loss) | |
| I0506 00:45:45.001518 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0465116 | |
| I0506 00:45:45.001530 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:45:45.001541 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:45:45.001554 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:45:45.001564 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:45:45.001575 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 00:45:45.001586 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 00:45:45.001597 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 00:45:45.001610 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:45:45.001621 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:45:45.001631 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:45:45.001642 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:45:45.001653 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:45:45.001664 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:45:45.001675 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:45:45.001687 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:45:45.001698 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:45:45.001708 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:45:45.001719 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:45:45.001730 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:45:45.001741 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:45:45.001752 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:45:45.001763 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:45:45.001775 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0506 00:45:45.001785 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.162791 | |
| I0506 00:45:45.001799 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.64507 (* 0.3 = 1.09352 loss) | |
| I0506 00:45:45.001812 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.27133 (* 0.3 = 0.381399 loss) | |
| I0506 00:45:45.001827 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.79537 (* 0.0272727 = 0.10351 loss) | |
| I0506 00:45:45.001839 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.69101 (* 0.0272727 = 0.100664 loss) | |
| I0506 00:45:45.001865 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.77383 (* 0.0272727 = 0.102923 loss) | |
| I0506 00:45:45.001879 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.39625 (* 0.0272727 = 0.0926251 loss) | |
| I0506 00:45:45.001893 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.23554 (* 0.0272727 = 0.088242 loss) | |
| I0506 00:45:45.001906 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.70087 (* 0.0272727 = 0.0736602 loss) | |
| I0506 00:45:45.001924 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.4751 (* 0.0272727 = 0.0402301 loss) | |
| I0506 00:45:45.001937 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.772073 (* 0.0272727 = 0.0210565 loss) | |
| I0506 00:45:45.001951 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0706693 (* 0.0272727 = 0.00192734 loss) | |
| I0506 00:45:45.001965 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0754229 (* 0.0272727 = 0.00205699 loss) | |
| I0506 00:45:45.001978 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0509024 (* 0.0272727 = 0.00138825 loss) | |
| I0506 00:45:45.001991 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0297053 (* 0.0272727 = 0.000810145 loss) | |
| I0506 00:45:45.002005 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0193039 (* 0.0272727 = 0.000526469 loss) | |
| I0506 00:45:45.002019 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0510748 (* 0.0272727 = 0.00139295 loss) | |
| I0506 00:45:45.002032 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0104123 (* 0.0272727 = 0.000283971 loss) | |
| I0506 00:45:45.002046 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00818533 (* 0.0272727 = 0.000223236 loss) | |
| I0506 00:45:45.002059 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00870661 (* 0.0272727 = 0.000237453 loss) | |
| I0506 00:45:45.002074 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0118304 (* 0.0272727 = 0.000322647 loss) | |
| I0506 00:45:45.002086 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00960166 (* 0.0272727 = 0.000261863 loss) | |
| I0506 00:45:45.002100 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00477042 (* 0.0272727 = 0.000130102 loss) | |
| I0506 00:45:45.002113 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0117106 (* 0.0272727 = 0.00031938 loss) | |
| I0506 00:45:45.002127 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00616618 (* 0.0272727 = 0.000168168 loss) | |
| I0506 00:45:45.002140 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0697674 | |
| I0506 00:45:45.002154 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 00:45:45.002167 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:45:45.002178 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:45:45.002189 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:45:45.002202 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 00:45:45.002213 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 00:45:45.002223 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 00:45:45.002235 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:45:45.002246 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:45:45.002259 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:45:45.002269 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:45:45.002280 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:45:45.002291 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:45:45.002302 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:45:45.002313 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:45:45.002324 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:45:45.002346 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:45:45.002359 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:45:45.002370 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:45:45.002382 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:45:45.002393 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:45:45.002403 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:45:45.002415 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.727273 | |
| I0506 00:45:45.002426 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.162791 | |
| I0506 00:45:45.002440 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.58659 (* 1 = 3.58659 loss) | |
| I0506 00:45:45.002454 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.18822 (* 1 = 1.18822 loss) | |
| I0506 00:45:45.002467 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.51287 (* 0.0909091 = 0.319352 loss) | |
| I0506 00:45:45.002480 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.87003 (* 0.0909091 = 0.351821 loss) | |
| I0506 00:45:45.002493 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.58333 (* 0.0909091 = 0.325758 loss) | |
| I0506 00:45:45.002507 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.79548 (* 0.0909091 = 0.254134 loss) | |
| I0506 00:45:45.002521 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.809 (* 0.0909091 = 0.255363 loss) | |
| I0506 00:45:45.002533 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.94261 (* 0.0909091 = 0.26751 loss) | |
| I0506 00:45:45.002547 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.17783 (* 0.0909091 = 0.107075 loss) | |
| I0506 00:45:45.002560 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.909811 (* 0.0909091 = 0.0827101 loss) | |
| I0506 00:45:45.002574 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0672546 (* 0.0909091 = 0.00611406 loss) | |
| I0506 00:45:45.002588 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0164414 (* 0.0909091 = 0.00149467 loss) | |
| I0506 00:45:45.002601 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.010662 (* 0.0909091 = 0.000969271 loss) | |
| I0506 00:45:45.002614 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00811591 (* 0.0909091 = 0.00073781 loss) | |
| I0506 00:45:45.002629 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00618807 (* 0.0909091 = 0.000562552 loss) | |
| I0506 00:45:45.002641 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00521302 (* 0.0909091 = 0.000473911 loss) | |
| I0506 00:45:45.002657 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00405041 (* 0.0909091 = 0.000368219 loss) | |
| I0506 00:45:45.002686 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0035341 (* 0.0909091 = 0.000321282 loss) | |
| I0506 00:45:45.002708 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00129096 (* 0.0909091 = 0.00011736 loss) | |
| I0506 00:45:45.002723 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0013747 (* 0.0909091 = 0.000124973 loss) | |
| I0506 00:45:45.002737 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000755556 (* 0.0909091 = 6.86869e-05 loss) | |
| I0506 00:45:45.002751 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000612622 (* 0.0909091 = 5.56929e-05 loss) | |
| I0506 00:45:45.002769 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000273954 (* 0.0909091 = 2.49049e-05 loss) | |
| I0506 00:45:45.002784 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000226074 (* 0.0909091 = 2.05522e-05 loss) | |
| I0506 00:45:45.002795 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:45:45.002807 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:45:45.002818 15760 solver.cpp:245] Train net output #149: total_confidence = 9.51512e-06 | |
| I0506 00:45:45.002852 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.14106e-05 | |
| I0506 00:45:45.002867 15760 sgd_solver.cpp:106] Iteration 6500, lr = 0.001 | |
| I0506 00:47:30.430546 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.2518 > 30) by scale factor 0.851019 | |
| I0506 00:47:32.301420 15760 solver.cpp:229] Iteration 7000, loss = 11.1788 | |
| I0506 00:47:32.301496 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0714286 | |
| I0506 00:47:32.301514 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:47:32.301527 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:47:32.301538 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:47:32.301550 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 00:47:32.301563 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 00:47:32.301574 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0506 00:47:32.301585 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 00:47:32.301597 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:47:32.301609 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:47:32.301621 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:47:32.301632 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:47:32.301643 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:47:32.301656 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:47:32.301666 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:47:32.301678 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:47:32.301690 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:47:32.301702 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:47:32.301713 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:47:32.301724 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:47:32.301736 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:47:32.301748 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:47:32.301759 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:47:32.301770 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.767045 | |
| I0506 00:47:32.301781 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.166667 | |
| I0506 00:47:32.301798 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.59231 (* 0.3 = 1.07769 loss) | |
| I0506 00:47:32.301812 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.05857 (* 0.3 = 0.31757 loss) | |
| I0506 00:47:32.301826 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.47736 (* 0.0272727 = 0.0948371 loss) | |
| I0506 00:47:32.301839 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.56611 (* 0.0272727 = 0.0972577 loss) | |
| I0506 00:47:32.301853 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.39115 (* 0.0272727 = 0.0924859 loss) | |
| I0506 00:47:32.301867 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.38379 (* 0.0272727 = 0.0922852 loss) | |
| I0506 00:47:32.301883 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.66915 (* 0.0272727 = 0.0727949 loss) | |
| I0506 00:47:32.301898 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.37804 (* 0.0272727 = 0.0648556 loss) | |
| I0506 00:47:32.301913 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.44082 (* 0.0272727 = 0.0392951 loss) | |
| I0506 00:47:32.301925 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.723215 (* 0.0272727 = 0.019724 loss) | |
| I0506 00:47:32.301940 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0423823 (* 0.0272727 = 0.00115588 loss) | |
| I0506 00:47:32.301954 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0432371 (* 0.0272727 = 0.00117919 loss) | |
| I0506 00:47:32.301967 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0213231 (* 0.0272727 = 0.000581539 loss) | |
| I0506 00:47:32.301981 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0431183 (* 0.0272727 = 0.00117595 loss) | |
| I0506 00:47:32.302027 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0276027 (* 0.0272727 = 0.000752801 loss) | |
| I0506 00:47:32.302043 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0199374 (* 0.0272727 = 0.000543746 loss) | |
| I0506 00:47:32.302057 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0130468 (* 0.0272727 = 0.000355821 loss) | |
| I0506 00:47:32.302070 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0113107 (* 0.0272727 = 0.000308474 loss) | |
| I0506 00:47:32.302084 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00645166 (* 0.0272727 = 0.000175954 loss) | |
| I0506 00:47:32.302098 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00602656 (* 0.0272727 = 0.000164361 loss) | |
| I0506 00:47:32.302111 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00877267 (* 0.0272727 = 0.000239255 loss) | |
| I0506 00:47:32.302125 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00747345 (* 0.0272727 = 0.000203821 loss) | |
| I0506 00:47:32.302139 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00679367 (* 0.0272727 = 0.000185282 loss) | |
| I0506 00:47:32.302152 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00621222 (* 0.0272727 = 0.000169424 loss) | |
| I0506 00:47:32.302165 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:47:32.302176 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:47:32.302187 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 00:47:32.302199 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 00:47:32.302211 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 00:47:32.302222 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 00:47:32.302233 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0506 00:47:32.302245 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 00:47:32.302256 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:47:32.302268 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:47:32.302279 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:47:32.302291 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:47:32.302304 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:47:32.302315 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:47:32.302326 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:47:32.302338 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:47:32.302350 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:47:32.302361 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:47:32.302371 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:47:32.302382 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:47:32.302393 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:47:32.302404 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:47:32.302415 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:47:32.302426 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0506 00:47:32.302438 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.047619 | |
| I0506 00:47:32.302453 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.59567 (* 0.3 = 1.0787 loss) | |
| I0506 00:47:32.302465 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.953823 (* 0.3 = 0.286147 loss) | |
| I0506 00:47:32.302479 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.35704 (* 0.0272727 = 0.0915558 loss) | |
| I0506 00:47:32.302494 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.52751 (* 0.0272727 = 0.096205 loss) | |
| I0506 00:47:32.302517 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.05925 (* 0.0272727 = 0.0834342 loss) | |
| I0506 00:47:32.302536 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.15636 (* 0.0272727 = 0.0860826 loss) | |
| I0506 00:47:32.302551 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.66404 (* 0.0272727 = 0.0726556 loss) | |
| I0506 00:47:32.302564 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.27067 (* 0.0272727 = 0.0619273 loss) | |
| I0506 00:47:32.302577 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.62327 (* 0.0272727 = 0.0442709 loss) | |
| I0506 00:47:32.302592 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.63151 (* 0.0272727 = 0.017223 loss) | |
| I0506 00:47:32.302605 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0274872 (* 0.0272727 = 0.000749651 loss) | |
| I0506 00:47:32.302619 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0207883 (* 0.0272727 = 0.000566954 loss) | |
| I0506 00:47:32.302633 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0115262 (* 0.0272727 = 0.00031435 loss) | |
| I0506 00:47:32.302646 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00993835 (* 0.0272727 = 0.000271046 loss) | |
| I0506 00:47:32.302660 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00785574 (* 0.0272727 = 0.000214247 loss) | |
| I0506 00:47:32.302673 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0137687 (* 0.0272727 = 0.000375509 loss) | |
| I0506 00:47:32.302687 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00712391 (* 0.0272727 = 0.000194288 loss) | |
| I0506 00:47:32.302701 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00391947 (* 0.0272727 = 0.000106895 loss) | |
| I0506 00:47:32.302714 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00551265 (* 0.0272727 = 0.000150345 loss) | |
| I0506 00:47:32.302728 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00457584 (* 0.0272727 = 0.000124796 loss) | |
| I0506 00:47:32.302742 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00168615 (* 0.0272727 = 4.59858e-05 loss) | |
| I0506 00:47:32.302755 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00218954 (* 0.0272727 = 5.97147e-05 loss) | |
| I0506 00:47:32.302769 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00285072 (* 0.0272727 = 7.77469e-05 loss) | |
| I0506 00:47:32.302783 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00357659 (* 0.0272727 = 9.75433e-05 loss) | |
| I0506 00:47:32.302795 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.047619 | |
| I0506 00:47:32.302806 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 00:47:32.302819 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:47:32.302830 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:47:32.302841 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0506 00:47:32.302853 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 00:47:32.302865 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 00:47:32.302875 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 00:47:32.302887 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:47:32.302899 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:47:32.302911 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:47:32.302922 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:47:32.302937 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:47:32.302948 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:47:32.302959 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:47:32.302970 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:47:32.302992 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:47:32.303004 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:47:32.303016 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:47:32.303027 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:47:32.303040 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:47:32.303050 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:47:32.303061 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:47:32.303072 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.772727 | |
| I0506 00:47:32.303084 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.238095 | |
| I0506 00:47:32.303098 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.00077 (* 1 = 3.00077 loss) | |
| I0506 00:47:32.303112 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.774452 (* 1 = 0.774452 loss) | |
| I0506 00:47:32.303125 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.41843 (* 0.0909091 = 0.310766 loss) | |
| I0506 00:47:32.303139 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.27046 (* 0.0909091 = 0.297315 loss) | |
| I0506 00:47:32.303153 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.10266 (* 0.0909091 = 0.28206 loss) | |
| I0506 00:47:32.303166 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.52203 (* 0.0909091 = 0.229275 loss) | |
| I0506 00:47:32.303179 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.37464 (* 0.0909091 = 0.215876 loss) | |
| I0506 00:47:32.303194 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.27004 (* 0.0909091 = 0.206367 loss) | |
| I0506 00:47:32.303206 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.10162 (* 0.0909091 = 0.100147 loss) | |
| I0506 00:47:32.303220 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.573018 (* 0.0909091 = 0.0520925 loss) | |
| I0506 00:47:32.303234 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00854008 (* 0.0909091 = 0.000776371 loss) | |
| I0506 00:47:32.303248 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00664499 (* 0.0909091 = 0.00060409 loss) | |
| I0506 00:47:32.303261 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00779798 (* 0.0909091 = 0.000708907 loss) | |
| I0506 00:47:32.303275 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00664851 (* 0.0909091 = 0.00060441 loss) | |
| I0506 00:47:32.303289 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00575081 (* 0.0909091 = 0.0005228 loss) | |
| I0506 00:47:32.303303 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00447502 (* 0.0909091 = 0.00040682 loss) | |
| I0506 00:47:32.303316 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00525202 (* 0.0909091 = 0.000477457 loss) | |
| I0506 00:47:32.303329 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00362054 (* 0.0909091 = 0.00032914 loss) | |
| I0506 00:47:32.303344 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00193277 (* 0.0909091 = 0.000175706 loss) | |
| I0506 00:47:32.303357 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00170616 (* 0.0909091 = 0.000155105 loss) | |
| I0506 00:47:32.303371 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00170603 (* 0.0909091 = 0.000155094 loss) | |
| I0506 00:47:32.303385 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0008655 (* 0.0909091 = 7.86818e-05 loss) | |
| I0506 00:47:32.303400 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00115464 (* 0.0909091 = 0.000104967 loss) | |
| I0506 00:47:32.303412 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000437421 (* 0.0909091 = 3.97655e-05 loss) | |
| I0506 00:47:32.303424 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:47:32.303436 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:47:32.303457 15760 solver.cpp:245] Train net output #149: total_confidence = 7.54778e-05 | |
| I0506 00:47:32.303469 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000720442 | |
| I0506 00:47:32.303483 15760 sgd_solver.cpp:106] Iteration 7000, lr = 0.001 | |
| I0506 00:48:15.384912 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.2339 > 30) by scale factor 0.902692 | |
| I0506 00:48:49.846098 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.5061 > 30) by scale factor 0.869412 | |
| I0506 00:48:56.059046 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0624 > 30) by scale factor 0.997925 | |
| I0506 00:48:57.992563 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.0538 > 30) by scale factor 0.880959 | |
| I0506 00:49:19.556100 15760 solver.cpp:229] Iteration 7500, loss = 11.0525 | |
| I0506 00:49:19.556170 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:49:19.556190 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:49:19.556201 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:49:19.556213 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:49:19.556224 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:49:19.556236 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:49:19.556248 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 00:49:19.556260 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 1 | |
| I0506 00:49:19.556272 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 00:49:19.556284 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:49:19.556295 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:49:19.556308 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:49:19.556318 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:49:19.556330 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:49:19.556342 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:49:19.556354 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:49:19.556365 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:49:19.556376 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:49:19.556388 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:49:19.556399 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:49:19.556411 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:49:19.556422 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:49:19.556434 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:49:19.556445 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.75 | |
| I0506 00:49:19.556457 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0681818 | |
| I0506 00:49:19.556473 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.78844 (* 0.3 = 1.13653 loss) | |
| I0506 00:49:19.556488 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.03195 (* 0.3 = 0.309584 loss) | |
| I0506 00:49:19.556502 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.57242 (* 0.0272727 = 0.0974298 loss) | |
| I0506 00:49:19.556517 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.06833 (* 0.0272727 = 0.110955 loss) | |
| I0506 00:49:19.556530 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.26139 (* 0.0272727 = 0.11622 loss) | |
| I0506 00:49:19.556545 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.92224 (* 0.0272727 = 0.10697 loss) | |
| I0506 00:49:19.556558 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.84452 (* 0.0272727 = 0.104851 loss) | |
| I0506 00:49:19.556571 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.49767 (* 0.0272727 = 0.0681184 loss) | |
| I0506 00:49:19.556586 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.259671 (* 0.0272727 = 0.00708193 loss) | |
| I0506 00:49:19.556599 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0537082 (* 0.0272727 = 0.00146477 loss) | |
| I0506 00:49:19.556614 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0276215 (* 0.0272727 = 0.000753313 loss) | |
| I0506 00:49:19.556628 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0178089 (* 0.0272727 = 0.000485698 loss) | |
| I0506 00:49:19.556680 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0214321 (* 0.0272727 = 0.000584513 loss) | |
| I0506 00:49:19.556696 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0217227 (* 0.0272727 = 0.000592438 loss) | |
| I0506 00:49:19.556710 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00804599 (* 0.0272727 = 0.000219436 loss) | |
| I0506 00:49:19.556723 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0240869 (* 0.0272727 = 0.000656915 loss) | |
| I0506 00:49:19.556737 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0255653 (* 0.0272727 = 0.000697235 loss) | |
| I0506 00:49:19.556751 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0208639 (* 0.0272727 = 0.000569014 loss) | |
| I0506 00:49:19.556766 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0212314 (* 0.0272727 = 0.000579037 loss) | |
| I0506 00:49:19.556779 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0223436 (* 0.0272727 = 0.000609372 loss) | |
| I0506 00:49:19.556793 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0166979 (* 0.0272727 = 0.000455398 loss) | |
| I0506 00:49:19.556807 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00866406 (* 0.0272727 = 0.000236292 loss) | |
| I0506 00:49:19.556821 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0124183 (* 0.0272727 = 0.00033868 loss) | |
| I0506 00:49:19.556835 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0124096 (* 0.0272727 = 0.000338443 loss) | |
| I0506 00:49:19.556848 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:49:19.556864 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:49:19.556875 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:49:19.556887 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:49:19.556900 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 00:49:19.556910 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 00:49:19.556922 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 00:49:19.556934 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 1 | |
| I0506 00:49:19.556946 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 00:49:19.556957 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:49:19.556967 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:49:19.556979 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:49:19.556990 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:49:19.557001 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:49:19.557013 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:49:19.557024 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:49:19.557035 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:49:19.557047 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:49:19.557059 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:49:19.557070 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:49:19.557081 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:49:19.557092 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:49:19.557103 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:49:19.557116 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0506 00:49:19.557144 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0681818 | |
| I0506 00:49:19.557160 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.86831 (* 0.3 = 1.16049 loss) | |
| I0506 00:49:19.557174 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.13594 (* 0.3 = 0.340782 loss) | |
| I0506 00:49:19.557201 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.44783 (* 0.0272727 = 0.0940316 loss) | |
| I0506 00:49:19.557216 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.95692 (* 0.0272727 = 0.107916 loss) | |
| I0506 00:49:19.557230 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 4.33617 (* 0.0272727 = 0.118259 loss) | |
| I0506 00:49:19.557245 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.20719 (* 0.0272727 = 0.114742 loss) | |
| I0506 00:49:19.557257 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.97049 (* 0.0272727 = 0.108286 loss) | |
| I0506 00:49:19.557271 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.91222 (* 0.0272727 = 0.0794242 loss) | |
| I0506 00:49:19.557284 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.284202 (* 0.0272727 = 0.00775097 loss) | |
| I0506 00:49:19.557298 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0997996 (* 0.0272727 = 0.00272181 loss) | |
| I0506 00:49:19.557312 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0244524 (* 0.0272727 = 0.000666885 loss) | |
| I0506 00:49:19.557327 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0208433 (* 0.0272727 = 0.000568453 loss) | |
| I0506 00:49:19.557340 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0200295 (* 0.0272727 = 0.00054626 loss) | |
| I0506 00:49:19.557354 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0174618 (* 0.0272727 = 0.000476232 loss) | |
| I0506 00:49:19.557368 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0343247 (* 0.0272727 = 0.000936129 loss) | |
| I0506 00:49:19.557382 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0167289 (* 0.0272727 = 0.000456244 loss) | |
| I0506 00:49:19.557395 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0162993 (* 0.0272727 = 0.000444526 loss) | |
| I0506 00:49:19.557410 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00941945 (* 0.0272727 = 0.000256894 loss) | |
| I0506 00:49:19.557422 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00570159 (* 0.0272727 = 0.000155498 loss) | |
| I0506 00:49:19.557436 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00922217 (* 0.0272727 = 0.000251514 loss) | |
| I0506 00:49:19.557451 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00979037 (* 0.0272727 = 0.00026701 loss) | |
| I0506 00:49:19.557464 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00751942 (* 0.0272727 = 0.000205075 loss) | |
| I0506 00:49:19.557477 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00429059 (* 0.0272727 = 0.000117016 loss) | |
| I0506 00:49:19.557492 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00765562 (* 0.0272727 = 0.00020879 loss) | |
| I0506 00:49:19.557503 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0227273 | |
| I0506 00:49:19.557515 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:49:19.557528 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:49:19.557538 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:49:19.557550 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 00:49:19.557561 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:49:19.557574 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 00:49:19.557585 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 1 | |
| I0506 00:49:19.557595 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 00:49:19.557607 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:49:19.557618 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:49:19.557629 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:49:19.557641 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:49:19.557652 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:49:19.557673 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:49:19.557685 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:49:19.557698 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:49:19.557708 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:49:19.557720 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:49:19.557728 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:49:19.557735 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:49:19.557742 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:49:19.557754 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:49:19.557765 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0506 00:49:19.557777 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.113636 | |
| I0506 00:49:19.557791 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.1283 (* 1 = 3.1283 loss) | |
| I0506 00:49:19.557804 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.901818 (* 1 = 0.901818 loss) | |
| I0506 00:49:19.557818 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.24575 (* 0.0909091 = 0.295068 loss) | |
| I0506 00:49:19.557832 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.11775 (* 0.0909091 = 0.283432 loss) | |
| I0506 00:49:19.557845 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.05924 (* 0.0909091 = 0.278113 loss) | |
| I0506 00:49:19.557858 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.35864 (* 0.0909091 = 0.305331 loss) | |
| I0506 00:49:19.557873 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.28031 (* 0.0909091 = 0.29821 loss) | |
| I0506 00:49:19.557885 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.60311 (* 0.0909091 = 0.236647 loss) | |
| I0506 00:49:19.557899 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.176115 (* 0.0909091 = 0.0160105 loss) | |
| I0506 00:49:19.557917 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0541397 (* 0.0909091 = 0.00492179 loss) | |
| I0506 00:49:19.557931 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00243154 (* 0.0909091 = 0.000221049 loss) | |
| I0506 00:49:19.557945 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00145145 (* 0.0909091 = 0.00013195 loss) | |
| I0506 00:49:19.557958 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00160723 (* 0.0909091 = 0.000146112 loss) | |
| I0506 00:49:19.557971 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00144612 (* 0.0909091 = 0.000131465 loss) | |
| I0506 00:49:19.557986 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00170219 (* 0.0909091 = 0.000154745 loss) | |
| I0506 00:49:19.557998 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00184661 (* 0.0909091 = 0.000167873 loss) | |
| I0506 00:49:19.558012 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00185095 (* 0.0909091 = 0.000168268 loss) | |
| I0506 00:49:19.558025 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00111364 (* 0.0909091 = 0.00010124 loss) | |
| I0506 00:49:19.558039 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000780794 (* 0.0909091 = 7.09813e-05 loss) | |
| I0506 00:49:19.558053 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00107941 (* 0.0909091 = 9.81286e-05 loss) | |
| I0506 00:49:19.558068 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000637976 (* 0.0909091 = 5.79978e-05 loss) | |
| I0506 00:49:19.558080 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000588501 (* 0.0909091 = 5.35001e-05 loss) | |
| I0506 00:49:19.558094 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00030625 (* 0.0909091 = 2.78409e-05 loss) | |
| I0506 00:49:19.558107 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000251835 (* 0.0909091 = 2.28941e-05 loss) | |
| I0506 00:49:19.558130 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:49:19.558141 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:49:19.558152 15760 solver.cpp:245] Train net output #149: total_confidence = 1.26082e-06 | |
| I0506 00:49:19.558164 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000213499 | |
| I0506 00:49:19.558182 15760 sgd_solver.cpp:106] Iteration 7500, lr = 0.001 | |
| I0506 00:50:10.922317 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.8727 > 30) by scale factor 0.912612 | |
| I0506 00:50:17.552284 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.5504 > 30) by scale factor 0.894178 | |
| I0506 00:51:06.867019 15760 solver.cpp:229] Iteration 8000, loss = 10.9567 | |
| I0506 00:51:06.867143 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0769231 | |
| I0506 00:51:06.867163 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:51:06.867177 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:51:06.867187 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 00:51:06.867199 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 00:51:06.867211 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 00:51:06.867223 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 00:51:06.867235 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 00:51:06.867246 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:51:06.867259 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 00:51:06.867270 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:51:06.867282 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:51:06.867295 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:51:06.867305 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:51:06.867317 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:51:06.867328 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:51:06.867341 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:51:06.867352 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:51:06.867363 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:51:06.867374 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:51:06.867385 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:51:06.867398 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:51:06.867408 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:51:06.867420 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.704545 | |
| I0506 00:51:06.867431 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.173077 | |
| I0506 00:51:06.867449 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.87185 (* 0.3 = 1.16156 loss) | |
| I0506 00:51:06.867462 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.37123 (* 0.3 = 0.411368 loss) | |
| I0506 00:51:06.867476 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.69045 (* 0.0272727 = 0.100649 loss) | |
| I0506 00:51:06.867491 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.56885 (* 0.0272727 = 0.124605 loss) | |
| I0506 00:51:06.867504 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.33216 (* 0.0272727 = 0.11815 loss) | |
| I0506 00:51:06.867518 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 4.32714 (* 0.0272727 = 0.118013 loss) | |
| I0506 00:51:06.867532 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.93024 (* 0.0272727 = 0.107188 loss) | |
| I0506 00:51:06.867545 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.26351 (* 0.0272727 = 0.0890048 loss) | |
| I0506 00:51:06.867559 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.47051 (* 0.0272727 = 0.0401049 loss) | |
| I0506 00:51:06.867573 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.631927 (* 0.0272727 = 0.0172344 loss) | |
| I0506 00:51:06.867586 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.636753 (* 0.0272727 = 0.017366 loss) | |
| I0506 00:51:06.867600 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.51555 (* 0.0272727 = 0.0140605 loss) | |
| I0506 00:51:06.867614 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.780591 (* 0.0272727 = 0.0212889 loss) | |
| I0506 00:51:06.867629 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.112546 (* 0.0272727 = 0.00306945 loss) | |
| I0506 00:51:06.867642 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.143977 (* 0.0272727 = 0.00392666 loss) | |
| I0506 00:51:06.867674 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.113084 (* 0.0272727 = 0.00308411 loss) | |
| I0506 00:51:06.867689 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0631169 (* 0.0272727 = 0.00172137 loss) | |
| I0506 00:51:06.867703 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0589932 (* 0.0272727 = 0.00160891 loss) | |
| I0506 00:51:06.867717 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0338548 (* 0.0272727 = 0.000923314 loss) | |
| I0506 00:51:06.867732 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.030605 (* 0.0272727 = 0.000834681 loss) | |
| I0506 00:51:06.867745 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0187678 (* 0.0272727 = 0.000511848 loss) | |
| I0506 00:51:06.867759 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.03252 (* 0.0272727 = 0.000886908 loss) | |
| I0506 00:51:06.867772 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0313437 (* 0.0272727 = 0.000854828 loss) | |
| I0506 00:51:06.867786 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0218382 (* 0.0272727 = 0.000595589 loss) | |
| I0506 00:51:06.867799 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0576923 | |
| I0506 00:51:06.867810 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:51:06.867821 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:51:06.867832 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:51:06.867843 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 00:51:06.867854 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0506 00:51:06.867866 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 00:51:06.867882 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 00:51:06.867889 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:51:06.867902 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 00:51:06.867914 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:51:06.867925 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:51:06.867936 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:51:06.867949 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:51:06.867959 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:51:06.867970 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:51:06.867981 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:51:06.867992 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:51:06.868003 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:51:06.868015 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:51:06.868026 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:51:06.868036 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:51:06.868047 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:51:06.868058 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0506 00:51:06.868070 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0961538 | |
| I0506 00:51:06.868083 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.70956 (* 0.3 = 1.11287 loss) | |
| I0506 00:51:06.868098 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.30027 (* 0.3 = 0.390082 loss) | |
| I0506 00:51:06.868111 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.5944 (* 0.0272727 = 0.098029 loss) | |
| I0506 00:51:06.868125 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.03767 (* 0.0272727 = 0.110118 loss) | |
| I0506 00:51:06.868149 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 4.46828 (* 0.0272727 = 0.121862 loss) | |
| I0506 00:51:06.868168 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.78126 (* 0.0272727 = 0.103125 loss) | |
| I0506 00:51:06.868182 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 4.04408 (* 0.0272727 = 0.110293 loss) | |
| I0506 00:51:06.868196 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.1447 (* 0.0272727 = 0.0857646 loss) | |
| I0506 00:51:06.868209 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.70878 (* 0.0272727 = 0.0466032 loss) | |
| I0506 00:51:06.868222 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.748104 (* 0.0272727 = 0.0204028 loss) | |
| I0506 00:51:06.868237 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.895461 (* 0.0272727 = 0.0244217 loss) | |
| I0506 00:51:06.868249 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.675607 (* 0.0272727 = 0.0184256 loss) | |
| I0506 00:51:06.868263 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.765761 (* 0.0272727 = 0.0208844 loss) | |
| I0506 00:51:06.868276 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.186477 (* 0.0272727 = 0.00508574 loss) | |
| I0506 00:51:06.868290 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.136768 (* 0.0272727 = 0.00373004 loss) | |
| I0506 00:51:06.868304 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.123565 (* 0.0272727 = 0.00336995 loss) | |
| I0506 00:51:06.868316 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0776778 (* 0.0272727 = 0.00211849 loss) | |
| I0506 00:51:06.868330 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0737898 (* 0.0272727 = 0.00201245 loss) | |
| I0506 00:51:06.868343 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0485096 (* 0.0272727 = 0.00132299 loss) | |
| I0506 00:51:06.868357 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0445539 (* 0.0272727 = 0.00121511 loss) | |
| I0506 00:51:06.868371 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0354874 (* 0.0272727 = 0.000967839 loss) | |
| I0506 00:51:06.868383 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0257701 (* 0.0272727 = 0.00070282 loss) | |
| I0506 00:51:06.868397 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0452421 (* 0.0272727 = 0.00123387 loss) | |
| I0506 00:51:06.868410 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0272014 (* 0.0272727 = 0.000741855 loss) | |
| I0506 00:51:06.868422 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0961538 | |
| I0506 00:51:06.868434 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:51:06.868445 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:51:06.868456 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:51:06.868468 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:51:06.868479 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 00:51:06.868490 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 00:51:06.868501 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 00:51:06.868512 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:51:06.868525 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 00:51:06.868535 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:51:06.868546 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:51:06.868557 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:51:06.868568 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:51:06.868579 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:51:06.868590 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:51:06.868602 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:51:06.868623 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:51:06.868635 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:51:06.868646 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:51:06.868657 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:51:06.868669 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:51:06.868680 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:51:06.868690 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0506 00:51:06.868702 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.192308 | |
| I0506 00:51:06.868716 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.47772 (* 1 = 3.47772 loss) | |
| I0506 00:51:06.868729 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.18621 (* 1 = 1.18621 loss) | |
| I0506 00:51:06.868742 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.17397 (* 0.0909091 = 0.288542 loss) | |
| I0506 00:51:06.868755 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.72822 (* 0.0909091 = 0.338929 loss) | |
| I0506 00:51:06.868768 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 4.10211 (* 0.0909091 = 0.37292 loss) | |
| I0506 00:51:06.868782 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.63457 (* 0.0909091 = 0.330416 loss) | |
| I0506 00:51:06.868795 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.14555 (* 0.0909091 = 0.285959 loss) | |
| I0506 00:51:06.868808 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.81035 (* 0.0909091 = 0.255486 loss) | |
| I0506 00:51:06.868823 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.17026 (* 0.0909091 = 0.106387 loss) | |
| I0506 00:51:06.868835 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.691949 (* 0.0909091 = 0.0629045 loss) | |
| I0506 00:51:06.868849 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.675912 (* 0.0909091 = 0.0614465 loss) | |
| I0506 00:51:06.868862 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.462631 (* 0.0909091 = 0.0420574 loss) | |
| I0506 00:51:06.868875 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.584726 (* 0.0909091 = 0.0531569 loss) | |
| I0506 00:51:06.868888 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.202504 (* 0.0909091 = 0.0184095 loss) | |
| I0506 00:51:06.868901 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.132688 (* 0.0909091 = 0.0120625 loss) | |
| I0506 00:51:06.868916 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.117889 (* 0.0909091 = 0.0107172 loss) | |
| I0506 00:51:06.868932 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0785487 (* 0.0909091 = 0.00714079 loss) | |
| I0506 00:51:06.868947 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0540603 (* 0.0909091 = 0.00491457 loss) | |
| I0506 00:51:06.868960 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0197566 (* 0.0909091 = 0.00179606 loss) | |
| I0506 00:51:06.868974 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0152736 (* 0.0909091 = 0.0013885 loss) | |
| I0506 00:51:06.868988 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00864989 (* 0.0909091 = 0.000786354 loss) | |
| I0506 00:51:06.869002 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00638144 (* 0.0909091 = 0.000580131 loss) | |
| I0506 00:51:06.869016 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00631912 (* 0.0909091 = 0.000574465 loss) | |
| I0506 00:51:06.869029 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00538587 (* 0.0909091 = 0.000489625 loss) | |
| I0506 00:51:06.869041 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:51:06.869052 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:51:06.869063 15760 solver.cpp:245] Train net output #149: total_confidence = 2.33091e-06 | |
| I0506 00:51:06.869086 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.61887e-05 | |
| I0506 00:51:06.869099 15760 sgd_solver.cpp:106] Iteration 8000, lr = 0.001 | |
| I0506 00:51:48.267562 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.9585 > 30) by scale factor 0.883431 | |
| I0506 00:52:48.641942 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 52.5746 > 30) by scale factor 0.570618 | |
| I0506 00:52:54.127571 15760 solver.cpp:229] Iteration 8500, loss = 10.7516 | |
| I0506 00:52:54.127640 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0425532 | |
| I0506 00:52:54.127657 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 00:52:54.127671 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:52:54.127684 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:52:54.127696 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 00:52:54.127708 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 00:52:54.127719 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 00:52:54.127732 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 00:52:54.127743 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 00:52:54.127755 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:52:54.127768 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:52:54.127779 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:52:54.127790 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:52:54.127802 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:52:54.127813 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:52:54.127825 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:52:54.127837 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:52:54.127849 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:52:54.127861 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:52:54.127873 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:52:54.127884 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:52:54.127899 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:52:54.127912 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:52:54.127923 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0506 00:52:54.127935 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.191489 | |
| I0506 00:52:54.127951 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.43153 (* 0.3 = 1.02946 loss) | |
| I0506 00:52:54.127966 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.985312 (* 0.3 = 0.295594 loss) | |
| I0506 00:52:54.127980 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.31198 (* 0.0272727 = 0.0903268 loss) | |
| I0506 00:52:54.127995 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.47487 (* 0.0272727 = 0.0947692 loss) | |
| I0506 00:52:54.128008 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.42879 (* 0.0272727 = 0.0935124 loss) | |
| I0506 00:52:54.128021 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.2696 (* 0.0272727 = 0.0891709 loss) | |
| I0506 00:52:54.128036 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.40413 (* 0.0272727 = 0.0928398 loss) | |
| I0506 00:52:54.128049 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.19551 (* 0.0272727 = 0.0871504 loss) | |
| I0506 00:52:54.128062 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.5863 (* 0.0272727 = 0.0705355 loss) | |
| I0506 00:52:54.128077 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.897786 (* 0.0272727 = 0.0244851 loss) | |
| I0506 00:52:54.128092 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.060299 (* 0.0272727 = 0.00164452 loss) | |
| I0506 00:52:54.128105 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0427422 (* 0.0272727 = 0.0011657 loss) | |
| I0506 00:52:54.128120 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0327025 (* 0.0272727 = 0.000891887 loss) | |
| I0506 00:52:54.128134 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0445189 (* 0.0272727 = 0.00121415 loss) | |
| I0506 00:52:54.128178 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0211145 (* 0.0272727 = 0.000575851 loss) | |
| I0506 00:52:54.128193 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0263237 (* 0.0272727 = 0.000717918 loss) | |
| I0506 00:52:54.128208 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.022586 (* 0.0272727 = 0.000615981 loss) | |
| I0506 00:52:54.128222 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0136389 (* 0.0272727 = 0.00037197 loss) | |
| I0506 00:52:54.128235 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0080343 (* 0.0272727 = 0.000219117 loss) | |
| I0506 00:52:54.128249 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00893449 (* 0.0272727 = 0.000243668 loss) | |
| I0506 00:52:54.128263 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00656172 (* 0.0272727 = 0.000178956 loss) | |
| I0506 00:52:54.128276 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00715192 (* 0.0272727 = 0.000195052 loss) | |
| I0506 00:52:54.128290 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00572702 (* 0.0272727 = 0.000156191 loss) | |
| I0506 00:52:54.128304 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00445453 (* 0.0272727 = 0.000121487 loss) | |
| I0506 00:52:54.128315 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0212766 | |
| I0506 00:52:54.128329 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:52:54.128340 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:52:54.128350 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 00:52:54.128362 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 00:52:54.128374 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 00:52:54.128386 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 00:52:54.128399 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 00:52:54.128407 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 00:52:54.128414 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:52:54.128427 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:52:54.128437 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:52:54.128453 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:52:54.128464 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:52:54.128475 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:52:54.128487 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:52:54.128499 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:52:54.128509 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:52:54.128520 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:52:54.128531 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:52:54.128542 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:52:54.128554 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:52:54.128566 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:52:54.128576 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.732955 | |
| I0506 00:52:54.128588 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.191489 | |
| I0506 00:52:54.128602 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.47952 (* 0.3 = 1.04386 loss) | |
| I0506 00:52:54.128615 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.02473 (* 0.3 = 0.30742 loss) | |
| I0506 00:52:54.128629 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.50966 (* 0.0272727 = 0.095718 loss) | |
| I0506 00:52:54.128643 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.38732 (* 0.0272727 = 0.0923816 loss) | |
| I0506 00:52:54.128667 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.40303 (* 0.0272727 = 0.09281 loss) | |
| I0506 00:52:54.128682 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.36017 (* 0.0272727 = 0.091641 loss) | |
| I0506 00:52:54.128695 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.2076 (* 0.0272727 = 0.0874801 loss) | |
| I0506 00:52:54.128710 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.24515 (* 0.0272727 = 0.0885041 loss) | |
| I0506 00:52:54.128723 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.49506 (* 0.0272727 = 0.0680471 loss) | |
| I0506 00:52:54.128736 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.734285 (* 0.0272727 = 0.020026 loss) | |
| I0506 00:52:54.128751 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.186903 (* 0.0272727 = 0.00509736 loss) | |
| I0506 00:52:54.128764 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.138347 (* 0.0272727 = 0.00377311 loss) | |
| I0506 00:52:54.128779 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.11638 (* 0.0272727 = 0.003174 loss) | |
| I0506 00:52:54.128793 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0712567 (* 0.0272727 = 0.00194337 loss) | |
| I0506 00:52:54.128806 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0693711 (* 0.0272727 = 0.00189194 loss) | |
| I0506 00:52:54.128819 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0453671 (* 0.0272727 = 0.00123728 loss) | |
| I0506 00:52:54.128834 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0509122 (* 0.0272727 = 0.00138851 loss) | |
| I0506 00:52:54.128846 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0225156 (* 0.0272727 = 0.000614061 loss) | |
| I0506 00:52:54.128860 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0268744 (* 0.0272727 = 0.000732939 loss) | |
| I0506 00:52:54.128873 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0158127 (* 0.0272727 = 0.000431256 loss) | |
| I0506 00:52:54.128886 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0156798 (* 0.0272727 = 0.000427632 loss) | |
| I0506 00:52:54.128901 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00753018 (* 0.0272727 = 0.000205369 loss) | |
| I0506 00:52:54.128913 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0142717 (* 0.0272727 = 0.000389228 loss) | |
| I0506 00:52:54.128927 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.013533 (* 0.0272727 = 0.000369081 loss) | |
| I0506 00:52:54.128939 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0638298 | |
| I0506 00:52:54.128954 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:52:54.128967 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0506 00:52:54.128978 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:52:54.128989 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 00:52:54.129000 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 00:52:54.129012 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 00:52:54.129024 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 00:52:54.129035 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:52:54.129046 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:52:54.129057 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:52:54.129070 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:52:54.129081 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:52:54.129091 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:52:54.129103 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:52:54.129114 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:52:54.129153 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:52:54.129166 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:52:54.129179 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:52:54.129189 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:52:54.129200 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:52:54.129211 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:52:54.129223 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:52:54.129235 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.738636 | |
| I0506 00:52:54.129245 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.319149 | |
| I0506 00:52:54.129259 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.91032 (* 1 = 2.91032 loss) | |
| I0506 00:52:54.129273 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.855718 (* 1 = 0.855718 loss) | |
| I0506 00:52:54.129287 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.50177 (* 0.0909091 = 0.227434 loss) | |
| I0506 00:52:54.129300 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.67085 (* 0.0909091 = 0.242805 loss) | |
| I0506 00:52:54.129314 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.17149 (* 0.0909091 = 0.288317 loss) | |
| I0506 00:52:54.129328 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.84966 (* 0.0909091 = 0.25906 loss) | |
| I0506 00:52:54.129340 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.8197 (* 0.0909091 = 0.256336 loss) | |
| I0506 00:52:54.129354 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.61965 (* 0.0909091 = 0.23815 loss) | |
| I0506 00:52:54.129367 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.33116 (* 0.0909091 = 0.211923 loss) | |
| I0506 00:52:54.129381 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.651906 (* 0.0909091 = 0.0592642 loss) | |
| I0506 00:52:54.129395 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0358951 (* 0.0909091 = 0.00326319 loss) | |
| I0506 00:52:54.129410 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0286475 (* 0.0909091 = 0.00260432 loss) | |
| I0506 00:52:54.129422 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0185314 (* 0.0909091 = 0.00168467 loss) | |
| I0506 00:52:54.129436 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00979562 (* 0.0909091 = 0.000890511 loss) | |
| I0506 00:52:54.129451 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0124579 (* 0.0909091 = 0.00113254 loss) | |
| I0506 00:52:54.129463 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00877838 (* 0.0909091 = 0.000798034 loss) | |
| I0506 00:52:54.129477 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00585565 (* 0.0909091 = 0.000532332 loss) | |
| I0506 00:52:54.129492 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00448248 (* 0.0909091 = 0.000407498 loss) | |
| I0506 00:52:54.129509 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0020351 (* 0.0909091 = 0.000185009 loss) | |
| I0506 00:52:54.129523 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0012843 (* 0.0909091 = 0.000116754 loss) | |
| I0506 00:52:54.129539 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00092513 (* 0.0909091 = 8.41027e-05 loss) | |
| I0506 00:52:54.129552 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000515956 (* 0.0909091 = 4.69051e-05 loss) | |
| I0506 00:52:54.129566 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000470173 (* 0.0909091 = 4.2743e-05 loss) | |
| I0506 00:52:54.129580 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000365104 (* 0.0909091 = 3.31913e-05 loss) | |
| I0506 00:52:54.129592 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:52:54.129603 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:52:54.129624 15760 solver.cpp:245] Train net output #149: total_confidence = 2.81988e-05 | |
| I0506 00:52:54.129637 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000177664 | |
| I0506 00:52:54.129650 15760 sgd_solver.cpp:106] Iteration 8500, lr = 0.001 | |
| I0506 00:53:27.546367 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.4644 > 30) by scale factor 0.760178 | |
| I0506 00:53:49.387737 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.3106 > 30) by scale factor 0.958142 | |
| I0506 00:54:41.152658 15760 solver.cpp:229] Iteration 9000, loss = 10.752 | |
| I0506 00:54:41.152775 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0526316 | |
| I0506 00:54:41.152796 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 00:54:41.152808 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:54:41.152820 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:54:41.152832 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 00:54:41.152843 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 00:54:41.152856 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0506 00:54:41.152868 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 00:54:41.152884 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 00:54:41.152895 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 00:54:41.152907 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 00:54:41.152920 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 00:54:41.152930 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:54:41.152945 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:54:41.152966 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:54:41.152981 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:54:41.152992 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:54:41.153003 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:54:41.153015 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:54:41.153026 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:54:41.153038 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:54:41.153049 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:54:41.153060 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:54:41.153072 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.789773 | |
| I0506 00:54:41.153084 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.210526 | |
| I0506 00:54:41.153100 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.27533 (* 0.3 = 1.2826 loss) | |
| I0506 00:54:41.153115 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.08086 (* 0.3 = 0.324257 loss) | |
| I0506 00:54:41.153144 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.62632 (* 0.0272727 = 0.126172 loss) | |
| I0506 00:54:41.153159 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 4.18196 (* 0.0272727 = 0.114053 loss) | |
| I0506 00:54:41.153173 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.13183 (* 0.0272727 = 0.112686 loss) | |
| I0506 00:54:41.153187 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 4.02435 (* 0.0272727 = 0.109755 loss) | |
| I0506 00:54:41.153201 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.1247 (* 0.0272727 = 0.0852191 loss) | |
| I0506 00:54:41.153216 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.16426 (* 0.0272727 = 0.0590252 loss) | |
| I0506 00:54:41.153230 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.50662 (* 0.0272727 = 0.0410897 loss) | |
| I0506 00:54:41.153254 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.519745 (* 0.0272727 = 0.0141749 loss) | |
| I0506 00:54:41.153270 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.273945 (* 0.0272727 = 0.00747123 loss) | |
| I0506 00:54:41.153285 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.234559 (* 0.0272727 = 0.00639706 loss) | |
| I0506 00:54:41.153298 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.216941 (* 0.0272727 = 0.00591656 loss) | |
| I0506 00:54:41.153312 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.175811 (* 0.0272727 = 0.00479484 loss) | |
| I0506 00:54:41.153326 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.139038 (* 0.0272727 = 0.00379194 loss) | |
| I0506 00:54:41.153357 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0938709 (* 0.0272727 = 0.00256011 loss) | |
| I0506 00:54:41.153373 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0841886 (* 0.0272727 = 0.00229605 loss) | |
| I0506 00:54:41.153388 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.101304 (* 0.0272727 = 0.00276284 loss) | |
| I0506 00:54:41.153400 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0947164 (* 0.0272727 = 0.00258318 loss) | |
| I0506 00:54:41.153414 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0940587 (* 0.0272727 = 0.00256524 loss) | |
| I0506 00:54:41.153429 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0533823 (* 0.0272727 = 0.00145588 loss) | |
| I0506 00:54:41.153442 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0550308 (* 0.0272727 = 0.00150084 loss) | |
| I0506 00:54:41.153455 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0464855 (* 0.0272727 = 0.00126779 loss) | |
| I0506 00:54:41.153470 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0605824 (* 0.0272727 = 0.00165225 loss) | |
| I0506 00:54:41.153481 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 00:54:41.153493 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:54:41.153506 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:54:41.153517 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 00:54:41.153528 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 00:54:41.153539 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 00:54:41.153551 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0506 00:54:41.153563 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 00:54:41.153574 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 00:54:41.153585 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 00:54:41.153596 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 00:54:41.153609 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 00:54:41.153620 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:54:41.153630 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:54:41.153641 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:54:41.153652 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:54:41.153664 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:54:41.153676 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:54:41.153687 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:54:41.153697 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:54:41.153709 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:54:41.153720 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:54:41.153730 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:54:41.153743 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.784091 | |
| I0506 00:54:41.153753 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.131579 | |
| I0506 00:54:41.153767 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.28226 (* 0.3 = 1.28468 loss) | |
| I0506 00:54:41.153780 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.00142 (* 0.3 = 0.300425 loss) | |
| I0506 00:54:41.153800 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.24048 (* 0.0272727 = 0.115649 loss) | |
| I0506 00:54:41.153815 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 4.22889 (* 0.0272727 = 0.115333 loss) | |
| I0506 00:54:41.153827 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.55691 (* 0.0272727 = 0.0970067 loss) | |
| I0506 00:54:41.153852 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.94358 (* 0.0272727 = 0.107552 loss) | |
| I0506 00:54:41.153867 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.08009 (* 0.0272727 = 0.0840025 loss) | |
| I0506 00:54:41.153882 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.12773 (* 0.0272727 = 0.058029 loss) | |
| I0506 00:54:41.153894 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.50334 (* 0.0272727 = 0.0410001 loss) | |
| I0506 00:54:41.153908 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.584234 (* 0.0272727 = 0.0159337 loss) | |
| I0506 00:54:41.153923 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.317162 (* 0.0272727 = 0.00864986 loss) | |
| I0506 00:54:41.153940 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.239492 (* 0.0272727 = 0.00653161 loss) | |
| I0506 00:54:41.153954 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.230786 (* 0.0272727 = 0.00629416 loss) | |
| I0506 00:54:41.153969 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.154348 (* 0.0272727 = 0.00420949 loss) | |
| I0506 00:54:41.153981 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.135166 (* 0.0272727 = 0.00368635 loss) | |
| I0506 00:54:41.153995 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.117124 (* 0.0272727 = 0.0031943 loss) | |
| I0506 00:54:41.154009 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.108009 (* 0.0272727 = 0.00294569 loss) | |
| I0506 00:54:41.154023 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0706505 (* 0.0272727 = 0.00192683 loss) | |
| I0506 00:54:41.154037 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0611958 (* 0.0272727 = 0.00166898 loss) | |
| I0506 00:54:41.154050 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0622615 (* 0.0272727 = 0.00169804 loss) | |
| I0506 00:54:41.154064 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.031097 (* 0.0272727 = 0.0008481 loss) | |
| I0506 00:54:41.154078 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0490017 (* 0.0272727 = 0.00133641 loss) | |
| I0506 00:54:41.154093 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0398314 (* 0.0272727 = 0.00108631 loss) | |
| I0506 00:54:41.154106 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0418978 (* 0.0272727 = 0.00114267 loss) | |
| I0506 00:54:41.154114 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0789474 | |
| I0506 00:54:41.154122 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 00:54:41.154136 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0506 00:54:41.154155 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:54:41.154171 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 00:54:41.154183 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 00:54:41.154196 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 00:54:41.154207 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0506 00:54:41.154218 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 00:54:41.154229 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 00:54:41.154242 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 00:54:41.154253 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 00:54:41.154263 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:54:41.154274 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:54:41.154285 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:54:41.154296 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:54:41.154306 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:54:41.154319 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:54:41.154340 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:54:41.154353 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:54:41.154364 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:54:41.154376 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:54:41.154386 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:54:41.154397 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.789773 | |
| I0506 00:54:41.154408 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.131579 | |
| I0506 00:54:41.154422 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.31501 (* 1 = 4.31501 loss) | |
| I0506 00:54:41.154435 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.06915 (* 1 = 1.06915 loss) | |
| I0506 00:54:41.154449 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.8728 (* 0.0909091 = 0.352073 loss) | |
| I0506 00:54:41.154463 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.54477 (* 0.0909091 = 0.322252 loss) | |
| I0506 00:54:41.154475 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.62764 (* 0.0909091 = 0.329786 loss) | |
| I0506 00:54:41.154489 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.84213 (* 0.0909091 = 0.349284 loss) | |
| I0506 00:54:41.154501 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.86781 (* 0.0909091 = 0.26071 loss) | |
| I0506 00:54:41.154515 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.82583 (* 0.0909091 = 0.165985 loss) | |
| I0506 00:54:41.154527 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.43346 (* 0.0909091 = 0.130314 loss) | |
| I0506 00:54:41.154541 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.318811 (* 0.0909091 = 0.0289828 loss) | |
| I0506 00:54:41.154554 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.111809 (* 0.0909091 = 0.0101644 loss) | |
| I0506 00:54:41.154568 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0856162 (* 0.0909091 = 0.00778329 loss) | |
| I0506 00:54:41.154582 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0682379 (* 0.0909091 = 0.00620345 loss) | |
| I0506 00:54:41.154595 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.049676 (* 0.0909091 = 0.004516 loss) | |
| I0506 00:54:41.154609 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0561225 (* 0.0909091 = 0.00510204 loss) | |
| I0506 00:54:41.154623 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0480109 (* 0.0909091 = 0.00436463 loss) | |
| I0506 00:54:41.154636 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0422868 (* 0.0909091 = 0.00384425 loss) | |
| I0506 00:54:41.154649 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0403587 (* 0.0909091 = 0.00366897 loss) | |
| I0506 00:54:41.154664 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0237668 (* 0.0909091 = 0.00216062 loss) | |
| I0506 00:54:41.154677 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0220263 (* 0.0909091 = 0.00200239 loss) | |
| I0506 00:54:41.154690 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0155641 (* 0.0909091 = 0.00141491 loss) | |
| I0506 00:54:41.154705 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0130077 (* 0.0909091 = 0.00118252 loss) | |
| I0506 00:54:41.154718 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0135548 (* 0.0909091 = 0.00123226 loss) | |
| I0506 00:54:41.154731 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00859373 (* 0.0909091 = 0.000781248 loss) | |
| I0506 00:54:41.154743 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:54:41.154754 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:54:41.154765 15760 solver.cpp:245] Train net output #149: total_confidence = 2.10098e-05 | |
| I0506 00:54:41.154777 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.04996e-05 | |
| I0506 00:54:41.154798 15760 sgd_solver.cpp:106] Iteration 9000, lr = 0.001 | |
| I0506 00:56:28.480145 15760 solver.cpp:229] Iteration 9500, loss = 10.6672 | |
| I0506 00:56:28.480332 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0333333 | |
| I0506 00:56:28.480355 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:56:28.480370 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 00:56:28.480381 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:56:28.480392 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 00:56:28.480404 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0506 00:56:28.480417 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 00:56:28.480428 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 00:56:28.480440 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 00:56:28.480453 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 00:56:28.480464 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0506 00:56:28.480476 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.75 | |
| I0506 00:56:28.480489 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 00:56:28.480500 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 00:56:28.480512 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 00:56:28.480525 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0506 00:56:28.480536 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:56:28.480548 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:56:28.480561 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:56:28.480572 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:56:28.480584 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:56:28.480595 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:56:28.480607 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:56:28.480618 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.670455 | |
| I0506 00:56:28.480630 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.183333 | |
| I0506 00:56:28.480648 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.74583 (* 0.3 = 1.12375 loss) | |
| I0506 00:56:28.480661 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.46244 (* 0.3 = 0.438732 loss) | |
| I0506 00:56:28.480675 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 4.01048 (* 0.0272727 = 0.109377 loss) | |
| I0506 00:56:28.480690 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.84825 (* 0.0272727 = 0.104952 loss) | |
| I0506 00:56:28.480703 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.20106 (* 0.0272727 = 0.114574 loss) | |
| I0506 00:56:28.480717 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.90272 (* 0.0272727 = 0.106438 loss) | |
| I0506 00:56:28.480731 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 4.51813 (* 0.0272727 = 0.123222 loss) | |
| I0506 00:56:28.480744 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.88911 (* 0.0272727 = 0.0787938 loss) | |
| I0506 00:56:28.480758 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.41712 (* 0.0272727 = 0.0659214 loss) | |
| I0506 00:56:28.480772 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.903247 (* 0.0272727 = 0.024634 loss) | |
| I0506 00:56:28.480787 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.30256 (* 0.0272727 = 0.0355244 loss) | |
| I0506 00:56:28.480799 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 1.31265 (* 0.0272727 = 0.0357995 loss) | |
| I0506 00:56:28.480813 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 1.57128 (* 0.0272727 = 0.042853 loss) | |
| I0506 00:56:28.480828 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.615826 (* 0.0272727 = 0.0167953 loss) | |
| I0506 00:56:28.480841 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.808543 (* 0.0272727 = 0.0220512 loss) | |
| I0506 00:56:28.480870 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.695929 (* 0.0272727 = 0.0189799 loss) | |
| I0506 00:56:28.480890 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.656389 (* 0.0272727 = 0.0179015 loss) | |
| I0506 00:56:28.480904 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0405641 (* 0.0272727 = 0.0011063 loss) | |
| I0506 00:56:28.480919 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0449023 (* 0.0272727 = 0.00122461 loss) | |
| I0506 00:56:28.480933 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0268221 (* 0.0272727 = 0.000731511 loss) | |
| I0506 00:56:28.480952 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.019339 (* 0.0272727 = 0.000527427 loss) | |
| I0506 00:56:28.480967 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.016643 (* 0.0272727 = 0.0004539 loss) | |
| I0506 00:56:28.480980 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0176316 (* 0.0272727 = 0.000480862 loss) | |
| I0506 00:56:28.480994 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0269962 (* 0.0272727 = 0.00073626 loss) | |
| I0506 00:56:28.481006 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0833333 | |
| I0506 00:56:28.481019 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 00:56:28.481030 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 00:56:28.481042 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:56:28.481053 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 00:56:28.481065 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0506 00:56:28.481076 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 00:56:28.481087 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 00:56:28.481098 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 00:56:28.481111 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 00:56:28.481143 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0506 00:56:28.481158 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.75 | |
| I0506 00:56:28.481169 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 00:56:28.481181 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 00:56:28.481194 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 00:56:28.481205 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0506 00:56:28.481216 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:56:28.481228 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:56:28.481240 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:56:28.481251 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:56:28.481262 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:56:28.481273 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:56:28.481284 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:56:28.481295 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.6875 | |
| I0506 00:56:28.481307 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.183333 | |
| I0506 00:56:28.481322 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.80969 (* 0.3 = 1.14291 loss) | |
| I0506 00:56:28.481341 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.47767 (* 0.3 = 0.443302 loss) | |
| I0506 00:56:28.481356 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 4.00865 (* 0.0272727 = 0.109327 loss) | |
| I0506 00:56:28.481369 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.95256 (* 0.0272727 = 0.107797 loss) | |
| I0506 00:56:28.481396 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.94329 (* 0.0272727 = 0.107544 loss) | |
| I0506 00:56:28.481411 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.5078 (* 0.0272727 = 0.12294 loss) | |
| I0506 00:56:28.481426 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 4.71646 (* 0.0272727 = 0.128631 loss) | |
| I0506 00:56:28.481438 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.89074 (* 0.0272727 = 0.0788385 loss) | |
| I0506 00:56:28.481452 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.01815 (* 0.0272727 = 0.0550405 loss) | |
| I0506 00:56:28.481465 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.905816 (* 0.0272727 = 0.0247041 loss) | |
| I0506 00:56:28.481479 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.44898 (* 0.0272727 = 0.0395176 loss) | |
| I0506 00:56:28.481492 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 1.0937 (* 0.0272727 = 0.0298282 loss) | |
| I0506 00:56:28.481505 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 1.3646 (* 0.0272727 = 0.0372163 loss) | |
| I0506 00:56:28.481519 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.730006 (* 0.0272727 = 0.0199092 loss) | |
| I0506 00:56:28.481533 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.823092 (* 0.0272727 = 0.022448 loss) | |
| I0506 00:56:28.481546 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.690946 (* 0.0272727 = 0.018844 loss) | |
| I0506 00:56:28.481559 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.809502 (* 0.0272727 = 0.0220773 loss) | |
| I0506 00:56:28.481573 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0350865 (* 0.0272727 = 0.000956903 loss) | |
| I0506 00:56:28.481586 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0249166 (* 0.0272727 = 0.000679544 loss) | |
| I0506 00:56:28.481600 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0343365 (* 0.0272727 = 0.000936449 loss) | |
| I0506 00:56:28.481614 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0174892 (* 0.0272727 = 0.000476979 loss) | |
| I0506 00:56:28.481627 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0201106 (* 0.0272727 = 0.000548472 loss) | |
| I0506 00:56:28.481642 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0221855 (* 0.0272727 = 0.000605058 loss) | |
| I0506 00:56:28.481654 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0241259 (* 0.0272727 = 0.000657979 loss) | |
| I0506 00:56:28.481667 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.05 | |
| I0506 00:56:28.481678 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 00:56:28.481690 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:56:28.481701 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:56:28.481712 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 00:56:28.481724 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0506 00:56:28.481735 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 00:56:28.481746 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 00:56:28.481757 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 00:56:28.481770 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 00:56:28.481781 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0506 00:56:28.481792 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.75 | |
| I0506 00:56:28.481803 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 00:56:28.481814 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 00:56:28.481827 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 00:56:28.481837 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0506 00:56:28.481848 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:56:28.481869 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:56:28.481883 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:56:28.481894 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:56:28.481905 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:56:28.481916 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:56:28.481930 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:56:28.481942 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.664773 | |
| I0506 00:56:28.481955 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.183333 | |
| I0506 00:56:28.481968 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.52624 (* 1 = 3.52624 loss) | |
| I0506 00:56:28.481981 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.36497 (* 1 = 1.36497 loss) | |
| I0506 00:56:28.481995 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.63245 (* 0.0909091 = 0.330223 loss) | |
| I0506 00:56:28.482008 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.48691 (* 0.0909091 = 0.316992 loss) | |
| I0506 00:56:28.482023 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.76068 (* 0.0909091 = 0.34188 loss) | |
| I0506 00:56:28.482035 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.61723 (* 0.0909091 = 0.328839 loss) | |
| I0506 00:56:28.482048 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.56412 (* 0.0909091 = 0.324011 loss) | |
| I0506 00:56:28.482062 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.45318 (* 0.0909091 = 0.223017 loss) | |
| I0506 00:56:28.482075 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.041 (* 0.0909091 = 0.185545 loss) | |
| I0506 00:56:28.482089 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.768896 (* 0.0909091 = 0.0698996 loss) | |
| I0506 00:56:28.482102 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.06431 (* 0.0909091 = 0.096755 loss) | |
| I0506 00:56:28.482115 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.892999 (* 0.0909091 = 0.0811817 loss) | |
| I0506 00:56:28.482130 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 1.06739 (* 0.0909091 = 0.0970357 loss) | |
| I0506 00:56:28.482142 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.388528 (* 0.0909091 = 0.0353208 loss) | |
| I0506 00:56:28.482156 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.598845 (* 0.0909091 = 0.0544404 loss) | |
| I0506 00:56:28.482167 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.461171 (* 0.0909091 = 0.0419246 loss) | |
| I0506 00:56:28.482182 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.402219 (* 0.0909091 = 0.0365653 loss) | |
| I0506 00:56:28.482197 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0779824 (* 0.0909091 = 0.00708931 loss) | |
| I0506 00:56:28.482210 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0384997 (* 0.0909091 = 0.00349997 loss) | |
| I0506 00:56:28.482223 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0197891 (* 0.0909091 = 0.00179901 loss) | |
| I0506 00:56:28.482237 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0157673 (* 0.0909091 = 0.00143339 loss) | |
| I0506 00:56:28.482250 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00989705 (* 0.0909091 = 0.000899732 loss) | |
| I0506 00:56:28.482264 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00632072 (* 0.0909091 = 0.000574611 loss) | |
| I0506 00:56:28.482277 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00946749 (* 0.0909091 = 0.000860681 loss) | |
| I0506 00:56:28.482290 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:56:28.482300 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:56:28.482311 15760 solver.cpp:245] Train net output #149: total_confidence = 5.63842e-07 | |
| I0506 00:56:28.482332 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 8.64962e-06 | |
| I0506 00:56:28.482347 15760 sgd_solver.cpp:106] Iteration 9500, lr = 0.001 | |
| I0506 00:57:15.484561 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.6454 > 30) by scale factor 0.738091 | |
| I0506 00:57:28.308847 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.9365 > 30) by scale factor 0.732843 | |
| I0506 00:58:15.612496 15760 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_10000.caffemodel | |
| I0506 00:58:16.070243 15760 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_10000.solverstate | |
| I0506 00:58:16.283087 15760 solver.cpp:338] Iteration 10000, Testing net (#0) | |
| I0506 00:58:52.263614 15760 solver.cpp:393] Test loss: 9.96764 | |
| I0506 00:58:52.263741 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0488992 | |
| I0506 00:58:52.263770 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.1 | |
| I0506 00:58:52.263793 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.083 | |
| I0506 00:58:52.263811 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.056 | |
| I0506 00:58:52.263823 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.157 | |
| I0506 00:58:52.263835 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.286 | |
| I0506 00:58:52.263846 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.449 | |
| I0506 00:58:52.263859 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.726 | |
| I0506 00:58:52.263870 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.91 | |
| I0506 00:58:52.263885 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.99 | |
| I0506 00:58:52.263896 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0506 00:58:52.263908 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0506 00:58:52.263919 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0506 00:58:52.263931 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0506 00:58:52.263942 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0506 00:58:52.263952 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0506 00:58:52.263963 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0506 00:58:52.263975 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0506 00:58:52.263986 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0506 00:58:52.263998 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0506 00:58:52.264008 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0506 00:58:52.264019 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 00:58:52.264031 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 00:58:52.264042 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.759683 | |
| I0506 00:58:52.264055 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.186877 | |
| I0506 00:58:52.264070 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.90343 (* 0.3 = 1.17103 loss) | |
| I0506 00:58:52.264083 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.03026 (* 0.3 = 0.309078 loss) | |
| I0506 00:58:52.264097 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.45167 (* 0.0272727 = 0.0941364 loss) | |
| I0506 00:58:52.264111 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.52966 (* 0.0272727 = 0.0962633 loss) | |
| I0506 00:58:52.264124 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.63291 (* 0.0272727 = 0.0990795 loss) | |
| I0506 00:58:52.264137 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.42171 (* 0.0272727 = 0.0933195 loss) | |
| I0506 00:58:52.264150 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 3.03831 (* 0.0272727 = 0.0828629 loss) | |
| I0506 00:58:52.264163 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.56934 (* 0.0272727 = 0.0700728 loss) | |
| I0506 00:58:52.264178 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.58666 (* 0.0272727 = 0.0432726 loss) | |
| I0506 00:58:52.264190 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 0.667311 (* 0.0272727 = 0.0181994 loss) | |
| I0506 00:58:52.264204 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.186906 (* 0.0272727 = 0.00509744 loss) | |
| I0506 00:58:52.264217 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.136748 (* 0.0272727 = 0.00372948 loss) | |
| I0506 00:58:52.264231 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.114558 (* 0.0272727 = 0.00312431 loss) | |
| I0506 00:58:52.264245 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0855736 (* 0.0272727 = 0.00233382 loss) | |
| I0506 00:58:52.264258 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0771584 (* 0.0272727 = 0.00210432 loss) | |
| I0506 00:58:52.264525 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0596202 (* 0.0272727 = 0.00162601 loss) | |
| I0506 00:58:52.264549 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0483694 (* 0.0272727 = 0.00131917 loss) | |
| I0506 00:58:52.264564 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.0361562 (* 0.0272727 = 0.000986078 loss) | |
| I0506 00:58:52.264577 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0283995 (* 0.0272727 = 0.000774533 loss) | |
| I0506 00:58:52.264592 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0258818 (* 0.0272727 = 0.000705868 loss) | |
| I0506 00:58:52.264606 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0178287 (* 0.0272727 = 0.000486237 loss) | |
| I0506 00:58:52.264621 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.022501 (* 0.0272727 = 0.000613663 loss) | |
| I0506 00:58:52.264634 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0205788 (* 0.0272727 = 0.000561241 loss) | |
| I0506 00:58:52.264649 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0182358 (* 0.0272727 = 0.000497341 loss) | |
| I0506 00:58:52.264662 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0534036 | |
| I0506 00:58:52.264673 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.107 | |
| I0506 00:58:52.264685 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.096 | |
| I0506 00:58:52.264696 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.079 | |
| I0506 00:58:52.264708 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.154 | |
| I0506 00:58:52.264719 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.279 | |
| I0506 00:58:52.264731 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.45 | |
| I0506 00:58:52.264742 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.726 | |
| I0506 00:58:52.264755 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.91 | |
| I0506 00:58:52.264765 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.99 | |
| I0506 00:58:52.264776 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0506 00:58:52.264788 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0506 00:58:52.264799 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0506 00:58:52.264811 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0506 00:58:52.264822 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0506 00:58:52.264832 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0506 00:58:52.264843 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0506 00:58:52.264854 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0506 00:58:52.264865 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0506 00:58:52.264879 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0506 00:58:52.264891 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0506 00:58:52.264902 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 00:58:52.264912 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 00:58:52.264924 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.760683 | |
| I0506 00:58:52.264935 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.200686 | |
| I0506 00:58:52.264950 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.77821 (* 0.3 = 1.13346 loss) | |
| I0506 00:58:52.264962 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.990767 (* 0.3 = 0.29723 loss) | |
| I0506 00:58:52.264976 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 3.3838 (* 0.0272727 = 0.0922853 loss) | |
| I0506 00:58:52.265002 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.43939 (* 0.0272727 = 0.0938016 loss) | |
| I0506 00:58:52.265019 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.5303 (* 0.0272727 = 0.0962809 loss) | |
| I0506 00:58:52.265048 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.36451 (* 0.0272727 = 0.0917593 loss) | |
| I0506 00:58:52.265063 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 3.01219 (* 0.0272727 = 0.0821507 loss) | |
| I0506 00:58:52.265076 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.56015 (* 0.0272727 = 0.0698222 loss) | |
| I0506 00:58:52.265089 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.58619 (* 0.0272727 = 0.0432596 loss) | |
| I0506 00:58:52.265103 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 0.688661 (* 0.0272727 = 0.0187817 loss) | |
| I0506 00:58:52.265138 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.204663 (* 0.0272727 = 0.00558171 loss) | |
| I0506 00:58:52.265156 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.157159 (* 0.0272727 = 0.00428615 loss) | |
| I0506 00:58:52.265171 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.122111 (* 0.0272727 = 0.00333031 loss) | |
| I0506 00:58:52.265184 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.101523 (* 0.0272727 = 0.0027688 loss) | |
| I0506 00:58:52.265197 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0858127 (* 0.0272727 = 0.00234035 loss) | |
| I0506 00:58:52.265211 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0683536 (* 0.0272727 = 0.00186419 loss) | |
| I0506 00:58:52.265224 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.0609621 (* 0.0272727 = 0.0016626 loss) | |
| I0506 00:58:52.265239 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.0431074 (* 0.0272727 = 0.00117566 loss) | |
| I0506 00:58:52.265251 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0269514 (* 0.0272727 = 0.000735037 loss) | |
| I0506 00:58:52.265265 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0257418 (* 0.0272727 = 0.000702048 loss) | |
| I0506 00:58:52.265278 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0222686 (* 0.0272727 = 0.000607326 loss) | |
| I0506 00:58:52.265292 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.021895 (* 0.0272727 = 0.000597136 loss) | |
| I0506 00:58:52.265305 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0203247 (* 0.0272727 = 0.00055431 loss) | |
| I0506 00:58:52.265319 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0206635 (* 0.0272727 = 0.00056355 loss) | |
| I0506 00:58:52.265331 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0808052 | |
| I0506 00:58:52.265342 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.133 | |
| I0506 00:58:52.265353 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.099 | |
| I0506 00:58:52.265365 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.082 | |
| I0506 00:58:52.265377 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.172 | |
| I0506 00:58:52.265388 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.306 | |
| I0506 00:58:52.265398 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.445 | |
| I0506 00:58:52.265410 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.726 | |
| I0506 00:58:52.265421 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.91 | |
| I0506 00:58:52.265432 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.99 | |
| I0506 00:58:52.265444 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.999 | |
| I0506 00:58:52.265455 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0506 00:58:52.265466 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0506 00:58:52.265477 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0506 00:58:52.265488 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0506 00:58:52.265502 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0506 00:58:52.265514 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0506 00:58:52.265525 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0506 00:58:52.265548 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0506 00:58:52.265561 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0506 00:58:52.265573 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0506 00:58:52.265583 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 00:58:52.265599 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 00:58:52.265619 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.762501 | |
| I0506 00:58:52.265635 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.235607 | |
| I0506 00:58:52.265648 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.14806 (* 1 = 3.14806 loss) | |
| I0506 00:58:52.265662 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.875052 (* 1 = 0.875052 loss) | |
| I0506 00:58:52.265676 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 2.97803 (* 0.0909091 = 0.27073 loss) | |
| I0506 00:58:52.265689 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 3.10629 (* 0.0909091 = 0.28239 loss) | |
| I0506 00:58:52.265702 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.21792 (* 0.0909091 = 0.292538 loss) | |
| I0506 00:58:52.265715 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 2.99786 (* 0.0909091 = 0.272532 loss) | |
| I0506 00:58:52.265728 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 2.63807 (* 0.0909091 = 0.239824 loss) | |
| I0506 00:58:52.265741 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.25463 (* 0.0909091 = 0.204967 loss) | |
| I0506 00:58:52.265754 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.37475 (* 0.0909091 = 0.124977 loss) | |
| I0506 00:58:52.265768 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.577035 (* 0.0909091 = 0.0524577 loss) | |
| I0506 00:58:52.265780 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.137019 (* 0.0909091 = 0.0124562 loss) | |
| I0506 00:58:52.265794 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0942238 (* 0.0909091 = 0.0085658 loss) | |
| I0506 00:58:52.265808 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.07254 (* 0.0909091 = 0.00659455 loss) | |
| I0506 00:58:52.265822 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0617776 (* 0.0909091 = 0.00561614 loss) | |
| I0506 00:58:52.265831 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0572792 (* 0.0909091 = 0.0052072 loss) | |
| I0506 00:58:52.265841 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.0452412 (* 0.0909091 = 0.00411283 loss) | |
| I0506 00:58:52.265856 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.0383829 (* 0.0909091 = 0.00348935 loss) | |
| I0506 00:58:52.265869 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.031392 (* 0.0909091 = 0.00285381 loss) | |
| I0506 00:58:52.265882 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.021861 (* 0.0909091 = 0.00198736 loss) | |
| I0506 00:58:52.265895 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0182784 (* 0.0909091 = 0.00166167 loss) | |
| I0506 00:58:52.265908 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0146292 (* 0.0909091 = 0.00132993 loss) | |
| I0506 00:58:52.265923 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0141077 (* 0.0909091 = 0.00128252 loss) | |
| I0506 00:58:52.265938 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0118563 (* 0.0909091 = 0.00107785 loss) | |
| I0506 00:58:52.265952 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0110283 (* 0.0909091 = 0.00100257 loss) | |
| I0506 00:58:52.265964 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 00:58:52.265974 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:58:52.265985 15760 solver.cpp:406] Test net output #149: total_confidence = 0.000378326 | |
| I0506 00:58:52.265996 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000485832 | |
| I0506 00:58:52.266019 15760 solver.cpp:338] Iteration 10000, Testing net (#1) | |
| I0506 00:59:28.166398 15760 solver.cpp:393] Test loss: 10.5642 | |
| I0506 00:59:28.166514 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0550997 | |
| I0506 00:59:28.166534 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.098 | |
| I0506 00:59:28.166548 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.088 | |
| I0506 00:59:28.166559 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.066 | |
| I0506 00:59:28.166571 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.152 | |
| I0506 00:59:28.166582 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.281 | |
| I0506 00:59:28.166595 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.404 | |
| I0506 00:59:28.166606 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.633 | |
| I0506 00:59:28.166618 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.792 | |
| I0506 00:59:28.166630 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.885 | |
| I0506 00:59:28.166642 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.911 | |
| I0506 00:59:28.166653 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.924 | |
| I0506 00:59:28.166666 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.936 | |
| I0506 00:59:28.166676 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.949 | |
| I0506 00:59:28.166688 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.96 | |
| I0506 00:59:28.166700 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.97 | |
| I0506 00:59:28.166712 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.979 | |
| I0506 00:59:28.166731 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.988 | |
| I0506 00:59:28.166743 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0506 00:59:28.166755 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0506 00:59:28.166767 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.999 | |
| I0506 00:59:28.166779 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 00:59:28.166790 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 00:59:28.166801 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.727683 | |
| I0506 00:59:28.166812 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.190298 | |
| I0506 00:59:28.166828 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.84116 (* 0.3 = 1.15235 loss) | |
| I0506 00:59:28.166842 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.15297 (* 0.3 = 0.345892 loss) | |
| I0506 00:59:28.166857 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.43357 (* 0.0272727 = 0.0936429 loss) | |
| I0506 00:59:28.166869 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.4968 (* 0.0272727 = 0.0953672 loss) | |
| I0506 00:59:28.166887 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.60228 (* 0.0272727 = 0.0982441 loss) | |
| I0506 00:59:28.166901 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.42712 (* 0.0272727 = 0.093467 loss) | |
| I0506 00:59:28.166914 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 3.04571 (* 0.0272727 = 0.0830648 loss) | |
| I0506 00:59:28.166929 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.67468 (* 0.0272727 = 0.0729457 loss) | |
| I0506 00:59:28.166941 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.85455 (* 0.0272727 = 0.0505786 loss) | |
| I0506 00:59:28.166954 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 1.07447 (* 0.0272727 = 0.0293038 loss) | |
| I0506 00:59:28.166968 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.602951 (* 0.0272727 = 0.0164441 loss) | |
| I0506 00:59:28.166981 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.51623 (* 0.0272727 = 0.014079 loss) | |
| I0506 00:59:28.166995 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.453426 (* 0.0272727 = 0.0123662 loss) | |
| I0506 00:59:28.167008 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.393031 (* 0.0272727 = 0.010719 loss) | |
| I0506 00:59:28.167021 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.332644 (* 0.0272727 = 0.0090721 loss) | |
| I0506 00:59:28.167055 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.274013 (* 0.0272727 = 0.00747309 loss) | |
| I0506 00:59:28.167070 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.213513 (* 0.0272727 = 0.00582309 loss) | |
| I0506 00:59:28.167089 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.158498 (* 0.0272727 = 0.00432266 loss) | |
| I0506 00:59:28.167104 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.102948 (* 0.0272727 = 0.00280767 loss) | |
| I0506 00:59:28.167119 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0520784 (* 0.0272727 = 0.00142032 loss) | |
| I0506 00:59:28.167131 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0395593 (* 0.0272727 = 0.00107889 loss) | |
| I0506 00:59:28.167145 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0274063 (* 0.0272727 = 0.000747445 loss) | |
| I0506 00:59:28.167158 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.018311 (* 0.0272727 = 0.000499391 loss) | |
| I0506 00:59:28.167171 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0163301 (* 0.0272727 = 0.000445366 loss) | |
| I0506 00:59:28.167183 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0575927 | |
| I0506 00:59:28.167196 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.112 | |
| I0506 00:59:28.167207 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.093 | |
| I0506 00:59:28.167218 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.077 | |
| I0506 00:59:28.167229 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.155 | |
| I0506 00:59:28.167240 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.282 | |
| I0506 00:59:28.167251 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.405 | |
| I0506 00:59:28.167263 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.633 | |
| I0506 00:59:28.167274 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.792 | |
| I0506 00:59:28.167285 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.885 | |
| I0506 00:59:28.167296 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.911 | |
| I0506 00:59:28.167309 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.924 | |
| I0506 00:59:28.167323 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.936 | |
| I0506 00:59:28.167335 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.949 | |
| I0506 00:59:28.167347 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.96 | |
| I0506 00:59:28.167358 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.97 | |
| I0506 00:59:28.167369 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.979 | |
| I0506 00:59:28.167381 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.988 | |
| I0506 00:59:28.167392 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0506 00:59:28.167403 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0506 00:59:28.167414 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.999 | |
| I0506 00:59:28.167426 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 00:59:28.167438 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 00:59:28.167448 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.728364 | |
| I0506 00:59:28.167459 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.207221 | |
| I0506 00:59:28.167472 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.72239 (* 0.3 = 1.11672 loss) | |
| I0506 00:59:28.167485 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.11036 (* 0.3 = 0.333108 loss) | |
| I0506 00:59:28.167500 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 3.36031 (* 0.0272727 = 0.0916449 loss) | |
| I0506 00:59:28.167512 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.41069 (* 0.0272727 = 0.0930188 loss) | |
| I0506 00:59:28.167537 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.49728 (* 0.0272727 = 0.0953802 loss) | |
| I0506 00:59:28.167552 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.37254 (* 0.0272727 = 0.0919784 loss) | |
| I0506 00:59:28.167565 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 3.01337 (* 0.0272727 = 0.0821827 loss) | |
| I0506 00:59:28.167578 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.65986 (* 0.0272727 = 0.0725416 loss) | |
| I0506 00:59:28.167592 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.85173 (* 0.0272727 = 0.0505017 loss) | |
| I0506 00:59:28.167604 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 1.08712 (* 0.0272727 = 0.0296488 loss) | |
| I0506 00:59:28.167618 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.622052 (* 0.0272727 = 0.016965 loss) | |
| I0506 00:59:28.167630 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.523458 (* 0.0272727 = 0.0142761 loss) | |
| I0506 00:59:28.167644 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.454983 (* 0.0272727 = 0.0124086 loss) | |
| I0506 00:59:28.167657 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.404688 (* 0.0272727 = 0.011037 loss) | |
| I0506 00:59:28.167670 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.332892 (* 0.0272727 = 0.00907887 loss) | |
| I0506 00:59:28.167685 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.277845 (* 0.0272727 = 0.0075776 loss) | |
| I0506 00:59:28.167697 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.219961 (* 0.0272727 = 0.00599894 loss) | |
| I0506 00:59:28.167711 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.167297 (* 0.0272727 = 0.00456265 loss) | |
| I0506 00:59:28.167724 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.104143 (* 0.0272727 = 0.00284026 loss) | |
| I0506 00:59:28.167737 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0511249 (* 0.0272727 = 0.00139431 loss) | |
| I0506 00:59:28.167752 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0430014 (* 0.0272727 = 0.00117276 loss) | |
| I0506 00:59:28.167764 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0277939 (* 0.0272727 = 0.000758016 loss) | |
| I0506 00:59:28.167778 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0205771 (* 0.0272727 = 0.000561193 loss) | |
| I0506 00:59:28.167790 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0196306 (* 0.0272727 = 0.000535381 loss) | |
| I0506 00:59:28.167803 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0831391 | |
| I0506 00:59:28.167814 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.121 | |
| I0506 00:59:28.167824 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.102 | |
| I0506 00:59:28.167835 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.103 | |
| I0506 00:59:28.167846 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.179 | |
| I0506 00:59:28.167858 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.322 | |
| I0506 00:59:28.167870 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.424 | |
| I0506 00:59:28.167881 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.636 | |
| I0506 00:59:28.167891 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.796 | |
| I0506 00:59:28.167902 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.888 | |
| I0506 00:59:28.167913 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.911 | |
| I0506 00:59:28.167927 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.924 | |
| I0506 00:59:28.167939 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.936 | |
| I0506 00:59:28.167951 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.949 | |
| I0506 00:59:28.167963 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.96 | |
| I0506 00:59:28.167973 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.97 | |
| I0506 00:59:28.167984 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.979 | |
| I0506 00:59:28.168006 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.988 | |
| I0506 00:59:28.168018 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0506 00:59:28.168030 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0506 00:59:28.168041 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.999 | |
| I0506 00:59:28.168052 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 00:59:28.168063 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 00:59:28.168074 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.731547 | |
| I0506 00:59:28.168086 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.241565 | |
| I0506 00:59:28.168099 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.16321 (* 1 = 3.16321 loss) | |
| I0506 00:59:28.168112 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.00049 (* 1 = 1.00049 loss) | |
| I0506 00:59:28.168125 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 2.9826 (* 0.0909091 = 0.271145 loss) | |
| I0506 00:59:28.168138 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 3.10081 (* 0.0909091 = 0.281891 loss) | |
| I0506 00:59:28.168151 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.19002 (* 0.0909091 = 0.290002 loss) | |
| I0506 00:59:28.168164 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 3.02565 (* 0.0909091 = 0.275059 loss) | |
| I0506 00:59:28.168177 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 2.63974 (* 0.0909091 = 0.239976 loss) | |
| I0506 00:59:28.168190 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.34779 (* 0.0909091 = 0.213436 loss) | |
| I0506 00:59:28.168203 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.64185 (* 0.0909091 = 0.149259 loss) | |
| I0506 00:59:28.168216 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.953071 (* 0.0909091 = 0.0866428 loss) | |
| I0506 00:59:28.168231 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.510933 (* 0.0909091 = 0.0464485 loss) | |
| I0506 00:59:28.168243 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.421544 (* 0.0909091 = 0.0383222 loss) | |
| I0506 00:59:28.168256 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.373685 (* 0.0909091 = 0.0339714 loss) | |
| I0506 00:59:28.168270 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.3338 (* 0.0909091 = 0.0303455 loss) | |
| I0506 00:59:28.168279 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.282813 (* 0.0909091 = 0.0257103 loss) | |
| I0506 00:59:28.168288 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.2333 (* 0.0909091 = 0.0212091 loss) | |
| I0506 00:59:28.168303 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.185442 (* 0.0909091 = 0.0168584 loss) | |
| I0506 00:59:28.168315 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.144675 (* 0.0909091 = 0.0131523 loss) | |
| I0506 00:59:28.168335 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0929149 (* 0.0909091 = 0.00844681 loss) | |
| I0506 00:59:28.168354 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0453755 (* 0.0909091 = 0.00412505 loss) | |
| I0506 00:59:28.168372 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0338278 (* 0.0909091 = 0.00307526 loss) | |
| I0506 00:59:28.168386 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.019295 (* 0.0909091 = 0.00175409 loss) | |
| I0506 00:59:28.168401 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00955671 (* 0.0909091 = 0.000868792 loss) | |
| I0506 00:59:28.168413 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00820083 (* 0.0909091 = 0.00074553 loss) | |
| I0506 00:59:28.168424 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 00:59:28.168436 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:59:28.168447 15760 solver.cpp:406] Test net output #149: total_confidence = 0.000421244 | |
| I0506 00:59:28.168468 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000417029 | |
| I0506 00:59:28.304327 15760 solver.cpp:229] Iteration 10000, loss = 10.6356 | |
| I0506 00:59:28.304385 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0506 00:59:28.304404 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 00:59:28.304416 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 00:59:28.304428 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 00:59:28.304440 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 00:59:28.304452 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 00:59:28.304463 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0506 00:59:28.304476 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 00:59:28.304487 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 00:59:28.304500 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 00:59:28.304512 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 00:59:28.304524 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 00:59:28.304536 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 00:59:28.304548 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 00:59:28.304559 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 00:59:28.304571 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 00:59:28.304582 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 00:59:28.304594 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 00:59:28.304605 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 00:59:28.304617 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 00:59:28.304628 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 00:59:28.304641 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 00:59:28.304651 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 00:59:28.304663 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.721591 | |
| I0506 00:59:28.304674 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.133333 | |
| I0506 00:59:28.304690 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.303 (* 0.3 = 0.9909 loss) | |
| I0506 00:59:28.304704 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.17134 (* 0.3 = 0.351402 loss) | |
| I0506 00:59:28.304718 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.77251 (* 0.0272727 = 0.102887 loss) | |
| I0506 00:59:28.304733 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.19829 (* 0.0272727 = 0.0872261 loss) | |
| I0506 00:59:28.304746 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.09528 (* 0.0272727 = 0.111689 loss) | |
| I0506 00:59:28.304760 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.13438 (* 0.0272727 = 0.0854831 loss) | |
| I0506 00:59:28.304774 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.42107 (* 0.0272727 = 0.0660292 loss) | |
| I0506 00:59:28.304787 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.0744 (* 0.0272727 = 0.0565745 loss) | |
| I0506 00:59:28.304805 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.50918 (* 0.0272727 = 0.0411596 loss) | |
| I0506 00:59:28.304821 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.06871 (* 0.0272727 = 0.0291465 loss) | |
| I0506 00:59:28.304834 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.00841 (* 0.0272727 = 0.0275021 loss) | |
| I0506 00:59:28.304848 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.621779 (* 0.0272727 = 0.0169576 loss) | |
| I0506 00:59:28.304862 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.842215 (* 0.0272727 = 0.0229695 loss) | |
| I0506 00:59:28.304906 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.296551 (* 0.0272727 = 0.00808774 loss) | |
| I0506 00:59:28.304921 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.307914 (* 0.0272727 = 0.00839764 loss) | |
| I0506 00:59:28.304936 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.193028 (* 0.0272727 = 0.0052644 loss) | |
| I0506 00:59:28.304950 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.238225 (* 0.0272727 = 0.00649705 loss) | |
| I0506 00:59:28.304965 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.195258 (* 0.0272727 = 0.00532521 loss) | |
| I0506 00:59:28.304978 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.122558 (* 0.0272727 = 0.0033425 loss) | |
| I0506 00:59:28.304992 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.070545 (* 0.0272727 = 0.00192395 loss) | |
| I0506 00:59:28.305006 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0639804 (* 0.0272727 = 0.00174492 loss) | |
| I0506 00:59:28.305021 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0658903 (* 0.0272727 = 0.00179701 loss) | |
| I0506 00:59:28.305033 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0711849 (* 0.0272727 = 0.00194141 loss) | |
| I0506 00:59:28.305047 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0481119 (* 0.0272727 = 0.00131214 loss) | |
| I0506 00:59:28.305059 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0444444 | |
| I0506 00:59:28.305071 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 00:59:28.305083 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0506 00:59:28.305095 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 00:59:28.305106 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 00:59:28.305131 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 00:59:28.305146 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0506 00:59:28.305158 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 00:59:28.305171 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 00:59:28.305181 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 00:59:28.305193 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 00:59:28.305205 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 00:59:28.305217 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 00:59:28.305228 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 00:59:28.305239 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 00:59:28.305250 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 00:59:28.305261 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 00:59:28.305272 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 00:59:28.305284 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 00:59:28.305295 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 00:59:28.305306 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 00:59:28.305318 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 00:59:28.305330 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 00:59:28.305341 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.732955 | |
| I0506 00:59:28.305352 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.266667 | |
| I0506 00:59:28.305366 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.19583 (* 0.3 = 0.958749 loss) | |
| I0506 00:59:28.305379 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.0916 (* 0.3 = 0.327479 loss) | |
| I0506 00:59:28.305393 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.70144 (* 0.0272727 = 0.100948 loss) | |
| I0506 00:59:28.305420 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.27555 (* 0.0272727 = 0.0893331 loss) | |
| I0506 00:59:28.305434 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.82602 (* 0.0272727 = 0.104346 loss) | |
| I0506 00:59:28.305449 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.29162 (* 0.0272727 = 0.0897715 loss) | |
| I0506 00:59:28.305462 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.28982 (* 0.0272727 = 0.0624496 loss) | |
| I0506 00:59:28.305476 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.6296 (* 0.0272727 = 0.0717163 loss) | |
| I0506 00:59:28.305490 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.27947 (* 0.0272727 = 0.0348946 loss) | |
| I0506 00:59:28.305502 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.919719 (* 0.0272727 = 0.0250833 loss) | |
| I0506 00:59:28.305516 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.14309 (* 0.0272727 = 0.0311753 loss) | |
| I0506 00:59:28.305529 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.465673 (* 0.0272727 = 0.0127002 loss) | |
| I0506 00:59:28.305543 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.647262 (* 0.0272727 = 0.0176526 loss) | |
| I0506 00:59:28.305557 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.170321 (* 0.0272727 = 0.00464512 loss) | |
| I0506 00:59:28.305570 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.170956 (* 0.0272727 = 0.00466242 loss) | |
| I0506 00:59:28.305584 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.17076 (* 0.0272727 = 0.00465709 loss) | |
| I0506 00:59:28.305598 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.112699 (* 0.0272727 = 0.00307361 loss) | |
| I0506 00:59:28.305611 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0604654 (* 0.0272727 = 0.00164906 loss) | |
| I0506 00:59:28.305625 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0283562 (* 0.0272727 = 0.000773351 loss) | |
| I0506 00:59:28.305639 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.037568 (* 0.0272727 = 0.00102458 loss) | |
| I0506 00:59:28.305652 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.03604 (* 0.0272727 = 0.00098291 loss) | |
| I0506 00:59:28.305666 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0240135 (* 0.0272727 = 0.000654915 loss) | |
| I0506 00:59:28.305680 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0303727 (* 0.0272727 = 0.000828347 loss) | |
| I0506 00:59:28.305693 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0347186 (* 0.0272727 = 0.000946872 loss) | |
| I0506 00:59:28.305706 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0222222 | |
| I0506 00:59:28.305718 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 00:59:28.305729 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 00:59:28.305740 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 00:59:28.305752 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 00:59:28.305763 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 00:59:28.305775 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 00:59:28.305788 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 00:59:28.305799 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 00:59:28.305809 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 00:59:28.305821 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 00:59:28.305833 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 00:59:28.305845 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 00:59:28.305861 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 00:59:28.305873 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 00:59:28.305891 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 00:59:28.305905 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 00:59:28.305917 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 00:59:28.305929 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 00:59:28.305943 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 00:59:28.305954 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 00:59:28.305965 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 00:59:28.305977 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 00:59:28.305989 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.727273 | |
| I0506 00:59:28.305999 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.244444 | |
| I0506 00:59:28.306013 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.07061 (* 1 = 3.07061 loss) | |
| I0506 00:59:28.306027 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.01035 (* 1 = 1.01035 loss) | |
| I0506 00:59:28.306041 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.25298 (* 0.0909091 = 0.295726 loss) | |
| I0506 00:59:28.306054 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.94275 (* 0.0909091 = 0.267523 loss) | |
| I0506 00:59:28.306067 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.45129 (* 0.0909091 = 0.313754 loss) | |
| I0506 00:59:28.306082 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.87947 (* 0.0909091 = 0.26177 loss) | |
| I0506 00:59:28.306094 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.36098 (* 0.0909091 = 0.214634 loss) | |
| I0506 00:59:28.306107 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.32278 (* 0.0909091 = 0.211162 loss) | |
| I0506 00:59:28.306120 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.16318 (* 0.0909091 = 0.105744 loss) | |
| I0506 00:59:28.306134 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.97473 (* 0.0909091 = 0.0886118 loss) | |
| I0506 00:59:28.306148 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.909218 (* 0.0909091 = 0.0826562 loss) | |
| I0506 00:59:28.306160 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.528028 (* 0.0909091 = 0.0480026 loss) | |
| I0506 00:59:28.306174 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.641086 (* 0.0909091 = 0.0582806 loss) | |
| I0506 00:59:28.306187 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.35663 (* 0.0909091 = 0.0324209 loss) | |
| I0506 00:59:28.306201 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.19882 (* 0.0909091 = 0.0180746 loss) | |
| I0506 00:59:28.306215 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.141375 (* 0.0909091 = 0.0128522 loss) | |
| I0506 00:59:28.306228 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.101308 (* 0.0909091 = 0.00920985 loss) | |
| I0506 00:59:28.306241 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0795631 (* 0.0909091 = 0.00723301 loss) | |
| I0506 00:59:28.306255 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0343418 (* 0.0909091 = 0.00312198 loss) | |
| I0506 00:59:28.306269 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.019265 (* 0.0909091 = 0.00175136 loss) | |
| I0506 00:59:28.306282 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0128025 (* 0.0909091 = 0.00116386 loss) | |
| I0506 00:59:28.306296 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0114918 (* 0.0909091 = 0.00104471 loss) | |
| I0506 00:59:28.306309 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00954312 (* 0.0909091 = 0.000867556 loss) | |
| I0506 00:59:28.306324 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00870823 (* 0.0909091 = 0.000791657 loss) | |
| I0506 00:59:28.306335 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 00:59:28.306346 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 00:59:28.306368 15760 solver.cpp:245] Train net output #149: total_confidence = 0.00015581 | |
| I0506 00:59:28.306381 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00075208 | |
| I0506 00:59:28.306394 15760 sgd_solver.cpp:106] Iteration 10000, lr = 0.001 | |
| I0506 00:59:53.840303 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.4995 > 30) by scale factor 0.895535 | |
| I0506 01:00:16.419589 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.5951 > 30) by scale factor 0.949514 | |
| I0506 01:01:15.483536 15760 solver.cpp:229] Iteration 10500, loss = 10.5483 | |
| I0506 01:01:15.483670 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.118644 | |
| I0506 01:01:15.483693 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:01:15.483706 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0506 01:01:15.483718 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:01:15.483731 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:01:15.483742 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 01:01:15.483754 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:01:15.483767 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:01:15.483778 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 01:01:15.483790 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 01:01:15.483803 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:01:15.483815 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:01:15.483827 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:01:15.483839 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 01:01:15.483850 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 01:01:15.483862 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0506 01:01:15.483876 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0506 01:01:15.483899 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0506 01:01:15.483917 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:01:15.483933 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:01:15.483945 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:01:15.483958 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:01:15.483969 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:01:15.483980 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.704545 | |
| I0506 01:01:15.483992 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.271186 | |
| I0506 01:01:15.484009 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.27133 (* 0.3 = 0.981398 loss) | |
| I0506 01:01:15.484024 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.23294 (* 0.3 = 0.369881 loss) | |
| I0506 01:01:15.484037 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.55605 (* 0.0272727 = 0.0969833 loss) | |
| I0506 01:01:15.484051 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.0148 (* 0.0272727 = 0.0822217 loss) | |
| I0506 01:01:15.484066 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.29544 (* 0.0272727 = 0.0898757 loss) | |
| I0506 01:01:15.484079 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.289 (* 0.0272727 = 0.0897 loss) | |
| I0506 01:01:15.484093 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.94715 (* 0.0272727 = 0.0803768 loss) | |
| I0506 01:01:15.484107 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.74355 (* 0.0272727 = 0.0748241 loss) | |
| I0506 01:01:15.484122 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.7012 (* 0.0272727 = 0.073669 loss) | |
| I0506 01:01:15.484135 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.77319 (* 0.0272727 = 0.0483597 loss) | |
| I0506 01:01:15.484149 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.31711 (* 0.0272727 = 0.0359211 loss) | |
| I0506 01:01:15.484163 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.557158 (* 0.0272727 = 0.0151952 loss) | |
| I0506 01:01:15.484177 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.767127 (* 0.0272727 = 0.0209216 loss) | |
| I0506 01:01:15.484191 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.650533 (* 0.0272727 = 0.0177418 loss) | |
| I0506 01:01:15.484223 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.692635 (* 0.0272727 = 0.0188901 loss) | |
| I0506 01:01:15.484238 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.706198 (* 0.0272727 = 0.0192599 loss) | |
| I0506 01:01:15.484252 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.942836 (* 0.0272727 = 0.0257137 loss) | |
| I0506 01:01:15.484267 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.849106 (* 0.0272727 = 0.0231574 loss) | |
| I0506 01:01:15.484282 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 1.04101 (* 0.0272727 = 0.0283911 loss) | |
| I0506 01:01:15.484297 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00598681 (* 0.0272727 = 0.000163277 loss) | |
| I0506 01:01:15.484310 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00630412 (* 0.0272727 = 0.000171931 loss) | |
| I0506 01:01:15.484324 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00370788 (* 0.0272727 = 0.000101124 loss) | |
| I0506 01:01:15.484338 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0033013 (* 0.0272727 = 9.00354e-05 loss) | |
| I0506 01:01:15.484354 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00447863 (* 0.0272727 = 0.000122145 loss) | |
| I0506 01:01:15.484365 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0847458 | |
| I0506 01:01:15.484377 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:01:15.484390 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:01:15.484401 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 01:01:15.484413 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:01:15.484424 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:01:15.484436 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:01:15.484448 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:01:15.484460 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0506 01:01:15.484472 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 01:01:15.484483 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:01:15.484495 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:01:15.484508 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:01:15.484519 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 01:01:15.484530 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 01:01:15.484542 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0506 01:01:15.484555 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0506 01:01:15.484565 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0506 01:01:15.484577 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:01:15.484589 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:01:15.484601 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:01:15.484612 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:01:15.484623 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:01:15.484635 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0506 01:01:15.484647 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.254237 | |
| I0506 01:01:15.484660 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.26079 (* 0.3 = 0.978236 loss) | |
| I0506 01:01:15.484674 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.16264 (* 0.3 = 0.348791 loss) | |
| I0506 01:01:15.484694 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.1686 (* 0.0272727 = 0.0864165 loss) | |
| I0506 01:01:15.484707 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.87479 (* 0.0272727 = 0.0784034 loss) | |
| I0506 01:01:15.484732 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.32295 (* 0.0272727 = 0.090626 loss) | |
| I0506 01:01:15.484747 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.46813 (* 0.0272727 = 0.0945854 loss) | |
| I0506 01:01:15.484761 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.48118 (* 0.0272727 = 0.0949413 loss) | |
| I0506 01:01:15.484776 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.74473 (* 0.0272727 = 0.0748562 loss) | |
| I0506 01:01:15.484789 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.78209 (* 0.0272727 = 0.0758752 loss) | |
| I0506 01:01:15.484802 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.67586 (* 0.0272727 = 0.0457052 loss) | |
| I0506 01:01:15.484817 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.09762 (* 0.0272727 = 0.0299352 loss) | |
| I0506 01:01:15.484830 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.621731 (* 0.0272727 = 0.0169563 loss) | |
| I0506 01:01:15.484844 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.649121 (* 0.0272727 = 0.0177033 loss) | |
| I0506 01:01:15.484858 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.875334 (* 0.0272727 = 0.0238728 loss) | |
| I0506 01:01:15.484871 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.593328 (* 0.0272727 = 0.0161817 loss) | |
| I0506 01:01:15.484885 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.6046 (* 0.0272727 = 0.0164891 loss) | |
| I0506 01:01:15.484899 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.834738 (* 0.0272727 = 0.0227656 loss) | |
| I0506 01:01:15.484912 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.959198 (* 0.0272727 = 0.0261599 loss) | |
| I0506 01:01:15.484928 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 1.03078 (* 0.0272727 = 0.0281123 loss) | |
| I0506 01:01:15.484943 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0186311 (* 0.0272727 = 0.000508121 loss) | |
| I0506 01:01:15.484958 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0219622 (* 0.0272727 = 0.00059897 loss) | |
| I0506 01:01:15.484972 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0196076 (* 0.0272727 = 0.000534752 loss) | |
| I0506 01:01:15.484985 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0179337 (* 0.0272727 = 0.000489102 loss) | |
| I0506 01:01:15.484999 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0140727 (* 0.0272727 = 0.000383802 loss) | |
| I0506 01:01:15.485011 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.135593 | |
| I0506 01:01:15.485023 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:01:15.485035 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0506 01:01:15.485046 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0506 01:01:15.485059 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:01:15.485070 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 01:01:15.485081 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:01:15.485093 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:01:15.485105 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0506 01:01:15.485128 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 01:01:15.485144 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:01:15.485157 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:01:15.485168 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:01:15.485180 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 01:01:15.485191 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 01:01:15.485203 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0506 01:01:15.485226 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0506 01:01:15.485239 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0506 01:01:15.485251 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:01:15.485263 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:01:15.485275 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:01:15.485286 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:01:15.485298 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:01:15.485309 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.710227 | |
| I0506 01:01:15.485321 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.355932 | |
| I0506 01:01:15.485335 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.03903 (* 1 = 3.03903 loss) | |
| I0506 01:01:15.485349 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.07905 (* 1 = 1.07905 loss) | |
| I0506 01:01:15.485363 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.40619 (* 0.0909091 = 0.309653 loss) | |
| I0506 01:01:15.485378 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.55486 (* 0.0909091 = 0.23226 loss) | |
| I0506 01:01:15.485390 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.99117 (* 0.0909091 = 0.271925 loss) | |
| I0506 01:01:15.485404 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.87439 (* 0.0909091 = 0.261308 loss) | |
| I0506 01:01:15.485417 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.21721 (* 0.0909091 = 0.201565 loss) | |
| I0506 01:01:15.485431 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.28533 (* 0.0909091 = 0.207757 loss) | |
| I0506 01:01:15.485445 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.33544 (* 0.0909091 = 0.212312 loss) | |
| I0506 01:01:15.485458 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.27318 (* 0.0909091 = 0.115744 loss) | |
| I0506 01:01:15.485472 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.771533 (* 0.0909091 = 0.0701394 loss) | |
| I0506 01:01:15.485486 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.433976 (* 0.0909091 = 0.0394524 loss) | |
| I0506 01:01:15.485499 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.505805 (* 0.0909091 = 0.0459823 loss) | |
| I0506 01:01:15.485517 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.531173 (* 0.0909091 = 0.0482884 loss) | |
| I0506 01:01:15.485532 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.398535 (* 0.0909091 = 0.0362305 loss) | |
| I0506 01:01:15.485546 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.559081 (* 0.0909091 = 0.0508255 loss) | |
| I0506 01:01:15.485559 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.865604 (* 0.0909091 = 0.0786913 loss) | |
| I0506 01:01:15.485574 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.708214 (* 0.0909091 = 0.0643831 loss) | |
| I0506 01:01:15.485586 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.91132 (* 0.0909091 = 0.0828473 loss) | |
| I0506 01:01:15.485600 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00361443 (* 0.0909091 = 0.000328584 loss) | |
| I0506 01:01:15.485615 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00239349 (* 0.0909091 = 0.00021759 loss) | |
| I0506 01:01:15.485628 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00245557 (* 0.0909091 = 0.000223234 loss) | |
| I0506 01:01:15.485641 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00142854 (* 0.0909091 = 0.000129867 loss) | |
| I0506 01:01:15.485656 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00157136 (* 0.0909091 = 0.000142851 loss) | |
| I0506 01:01:15.485667 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:01:15.485678 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:01:15.485689 15760 solver.cpp:245] Train net output #149: total_confidence = 7.32735e-05 | |
| I0506 01:01:15.485710 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000747179 | |
| I0506 01:01:15.485724 15760 sgd_solver.cpp:106] Iteration 10500, lr = 0.001 | |
| I0506 01:02:19.659266 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.3831 > 30) by scale factor 0.926409 | |
| I0506 01:03:02.706852 15760 solver.cpp:229] Iteration 11000, loss = 10.4638 | |
| I0506 01:03:02.706982 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0384615 | |
| I0506 01:03:02.707003 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:03:02.707016 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0506 01:03:02.707028 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:03:02.707041 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:03:02.707052 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 01:03:02.707064 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 01:03:02.707077 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:03:02.707088 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:03:02.707100 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:03:02.707114 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:03:02.707125 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:03:02.707137 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:03:02.707149 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:03:02.707160 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:03:02.707172 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:03:02.707185 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:03:02.707196 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:03:02.707207 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:03:02.707219 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:03:02.707231 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:03:02.707242 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:03:02.707253 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:03:02.707265 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0506 01:03:02.707278 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.173077 | |
| I0506 01:03:02.707293 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.36283 (* 0.3 = 1.00885 loss) | |
| I0506 01:03:02.707307 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.18883 (* 0.3 = 0.35665 loss) | |
| I0506 01:03:02.707321 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.90145 (* 0.0272727 = 0.0791305 loss) | |
| I0506 01:03:02.707335 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.31946 (* 0.0272727 = 0.0905306 loss) | |
| I0506 01:03:02.707350 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.42457 (* 0.0272727 = 0.0933974 loss) | |
| I0506 01:03:02.707363 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.77122 (* 0.0272727 = 0.102851 loss) | |
| I0506 01:03:02.707377 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.52534 (* 0.0272727 = 0.0961455 loss) | |
| I0506 01:03:02.707391 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.48366 (* 0.0272727 = 0.0950089 loss) | |
| I0506 01:03:02.707404 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.805 (* 0.0272727 = 0.0764999 loss) | |
| I0506 01:03:02.707418 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.7063 (* 0.0272727 = 0.0465354 loss) | |
| I0506 01:03:02.707432 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.486952 (* 0.0272727 = 0.0132805 loss) | |
| I0506 01:03:02.707448 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.105908 (* 0.0272727 = 0.00288839 loss) | |
| I0506 01:03:02.707461 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.113852 (* 0.0272727 = 0.00310506 loss) | |
| I0506 01:03:02.707475 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0713797 (* 0.0272727 = 0.00194672 loss) | |
| I0506 01:03:02.707489 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.050942 (* 0.0272727 = 0.00138933 loss) | |
| I0506 01:03:02.707522 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.05799 (* 0.0272727 = 0.00158155 loss) | |
| I0506 01:03:02.707537 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0362265 (* 0.0272727 = 0.000987994 loss) | |
| I0506 01:03:02.707552 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.019965 (* 0.0272727 = 0.000544501 loss) | |
| I0506 01:03:02.707566 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00948847 (* 0.0272727 = 0.000258777 loss) | |
| I0506 01:03:02.707581 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00668053 (* 0.0272727 = 0.000182196 loss) | |
| I0506 01:03:02.707594 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00599524 (* 0.0272727 = 0.000163506 loss) | |
| I0506 01:03:02.707608 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00617411 (* 0.0272727 = 0.000168385 loss) | |
| I0506 01:03:02.707623 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00284744 (* 0.0272727 = 7.76576e-05 loss) | |
| I0506 01:03:02.707636 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00629933 (* 0.0272727 = 0.0001718 loss) | |
| I0506 01:03:02.707648 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0576923 | |
| I0506 01:03:02.707661 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.5 | |
| I0506 01:03:02.707674 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:03:02.707685 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 01:03:02.707696 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:03:02.707708 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 01:03:02.707720 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 01:03:02.707731 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:03:02.707743 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:03:02.707756 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:03:02.707767 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:03:02.707778 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:03:02.707789 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:03:02.707800 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:03:02.707811 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:03:02.707823 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:03:02.707834 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:03:02.707845 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:03:02.707856 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:03:02.707869 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:03:02.707883 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:03:02.707895 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:03:02.707906 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:03:02.707918 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0506 01:03:02.707931 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.173077 | |
| I0506 01:03:02.707943 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.22953 (* 0.3 = 0.96886 loss) | |
| I0506 01:03:02.707957 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.25325 (* 0.3 = 0.375975 loss) | |
| I0506 01:03:02.707972 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.33298 (* 0.0272727 = 0.0636266 loss) | |
| I0506 01:03:02.707985 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.48282 (* 0.0272727 = 0.0949861 loss) | |
| I0506 01:03:02.708014 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.36083 (* 0.0272727 = 0.091659 loss) | |
| I0506 01:03:02.708030 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.78153 (* 0.0272727 = 0.103133 loss) | |
| I0506 01:03:02.708044 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.3376 (* 0.0272727 = 0.0910255 loss) | |
| I0506 01:03:02.708057 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.7032 (* 0.0272727 = 0.100996 loss) | |
| I0506 01:03:02.708071 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.50497 (* 0.0272727 = 0.0683173 loss) | |
| I0506 01:03:02.708086 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.55227 (* 0.0272727 = 0.0423347 loss) | |
| I0506 01:03:02.708099 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.505009 (* 0.0272727 = 0.013773 loss) | |
| I0506 01:03:02.708113 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.106214 (* 0.0272727 = 0.00289675 loss) | |
| I0506 01:03:02.708127 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0810997 (* 0.0272727 = 0.00221181 loss) | |
| I0506 01:03:02.708142 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0536257 (* 0.0272727 = 0.00146252 loss) | |
| I0506 01:03:02.708155 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0256581 (* 0.0272727 = 0.000699766 loss) | |
| I0506 01:03:02.708168 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0263094 (* 0.0272727 = 0.000717528 loss) | |
| I0506 01:03:02.708183 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.028855 (* 0.0272727 = 0.000786956 loss) | |
| I0506 01:03:02.708196 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0124511 (* 0.0272727 = 0.000339576 loss) | |
| I0506 01:03:02.708210 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0142428 (* 0.0272727 = 0.00038844 loss) | |
| I0506 01:03:02.708225 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00872765 (* 0.0272727 = 0.000238027 loss) | |
| I0506 01:03:02.708238 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00559468 (* 0.0272727 = 0.000152582 loss) | |
| I0506 01:03:02.708252 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00299132 (* 0.0272727 = 8.15816e-05 loss) | |
| I0506 01:03:02.708266 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00267565 (* 0.0272727 = 7.29723e-05 loss) | |
| I0506 01:03:02.708279 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00370963 (* 0.0272727 = 0.000101172 loss) | |
| I0506 01:03:02.708292 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.134615 | |
| I0506 01:03:02.708303 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:03:02.708314 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:03:02.708326 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:03:02.708338 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:03:02.708349 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 01:03:02.708361 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0506 01:03:02.708372 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:03:02.708384 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:03:02.708395 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:03:02.708407 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:03:02.708418 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:03:02.708430 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:03:02.708441 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:03:02.708451 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:03:02.708463 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:03:02.708474 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:03:02.708495 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:03:02.708508 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:03:02.708520 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:03:02.708531 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:03:02.708542 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:03:02.708554 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:03:02.708565 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0506 01:03:02.708576 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.25 | |
| I0506 01:03:02.708590 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.07338 (* 1 = 3.07338 loss) | |
| I0506 01:03:02.708603 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.0299 (* 1 = 1.0299 loss) | |
| I0506 01:03:02.708617 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.25936 (* 0.0909091 = 0.205396 loss) | |
| I0506 01:03:02.708631 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.11415 (* 0.0909091 = 0.283105 loss) | |
| I0506 01:03:02.708644 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.02855 (* 0.0909091 = 0.275322 loss) | |
| I0506 01:03:02.708658 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.52432 (* 0.0909091 = 0.320393 loss) | |
| I0506 01:03:02.708672 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.06209 (* 0.0909091 = 0.278371 loss) | |
| I0506 01:03:02.708691 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.02361 (* 0.0909091 = 0.274874 loss) | |
| I0506 01:03:02.708701 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.29617 (* 0.0909091 = 0.208742 loss) | |
| I0506 01:03:02.708715 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.22708 (* 0.0909091 = 0.111553 loss) | |
| I0506 01:03:02.708729 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.450954 (* 0.0909091 = 0.0409958 loss) | |
| I0506 01:03:02.708742 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0716993 (* 0.0909091 = 0.00651812 loss) | |
| I0506 01:03:02.708755 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0493934 (* 0.0909091 = 0.00449031 loss) | |
| I0506 01:03:02.708770 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0217353 (* 0.0909091 = 0.00197594 loss) | |
| I0506 01:03:02.708783 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0154308 (* 0.0909091 = 0.0014028 loss) | |
| I0506 01:03:02.708796 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00819239 (* 0.0909091 = 0.000744763 loss) | |
| I0506 01:03:02.708811 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00275524 (* 0.0909091 = 0.000250476 loss) | |
| I0506 01:03:02.708823 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00245053 (* 0.0909091 = 0.000222776 loss) | |
| I0506 01:03:02.708837 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000512455 (* 0.0909091 = 4.65868e-05 loss) | |
| I0506 01:03:02.708852 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000455629 (* 0.0909091 = 4.14208e-05 loss) | |
| I0506 01:03:02.708864 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000244467 (* 0.0909091 = 2.22243e-05 loss) | |
| I0506 01:03:02.708878 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 9.52487e-05 (* 0.0909091 = 8.65897e-06 loss) | |
| I0506 01:03:02.708892 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 8.20351e-05 (* 0.0909091 = 7.45774e-06 loss) | |
| I0506 01:03:02.708905 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000168938 (* 0.0909091 = 1.5358e-05 loss) | |
| I0506 01:03:02.708917 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:03:02.708931 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:03:02.708943 15760 solver.cpp:245] Train net output #149: total_confidence = 8.2572e-07 | |
| I0506 01:03:02.708966 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.11067e-06 | |
| I0506 01:03:02.708979 15760 sgd_solver.cpp:106] Iteration 11000, lr = 0.001 | |
| I0506 01:03:20.575698 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.0641 > 30) by scale factor 0.855575 | |
| I0506 01:04:49.923508 15760 solver.cpp:229] Iteration 11500, loss = 10.4523 | |
| I0506 01:04:49.923650 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0208333 | |
| I0506 01:04:49.923671 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:04:49.923684 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:04:49.923696 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:04:49.923708 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:04:49.923720 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:04:49.923732 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 01:04:49.923744 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 01:04:49.923755 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:04:49.923768 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:04:49.923779 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:04:49.923791 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:04:49.923804 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:04:49.923815 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 01:04:49.923827 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 01:04:49.923840 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0506 01:04:49.923852 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:04:49.923864 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:04:49.923882 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:04:49.923893 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:04:49.923905 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:04:49.923918 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:04:49.923928 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:04:49.923940 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0506 01:04:49.923952 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.208333 | |
| I0506 01:04:49.923969 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.25259 (* 0.3 = 0.975776 loss) | |
| I0506 01:04:49.923991 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.00478 (* 0.3 = 0.301433 loss) | |
| I0506 01:04:49.924006 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.30956 (* 0.0272727 = 0.0902608 loss) | |
| I0506 01:04:49.924021 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.36205 (* 0.0272727 = 0.0916922 loss) | |
| I0506 01:04:49.924034 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.6784 (* 0.0272727 = 0.10032 loss) | |
| I0506 01:04:49.924048 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.97611 (* 0.0272727 = 0.0811667 loss) | |
| I0506 01:04:49.924062 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.29914 (* 0.0272727 = 0.0627037 loss) | |
| I0506 01:04:49.924077 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.75434 (* 0.0272727 = 0.0478456 loss) | |
| I0506 01:04:49.924089 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.40739 (* 0.0272727 = 0.0383835 loss) | |
| I0506 01:04:49.924103 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.916214 (* 0.0272727 = 0.0249877 loss) | |
| I0506 01:04:49.924116 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.472512 (* 0.0272727 = 0.0128867 loss) | |
| I0506 01:04:49.924130 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.507841 (* 0.0272727 = 0.0138502 loss) | |
| I0506 01:04:49.924144 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.566653 (* 0.0272727 = 0.0154542 loss) | |
| I0506 01:04:49.924159 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.628099 (* 0.0272727 = 0.01713 loss) | |
| I0506 01:04:49.924190 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.58875 (* 0.0272727 = 0.0160568 loss) | |
| I0506 01:04:49.924206 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.610776 (* 0.0272727 = 0.0166575 loss) | |
| I0506 01:04:49.924219 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.507857 (* 0.0272727 = 0.0138507 loss) | |
| I0506 01:04:49.924234 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0432954 (* 0.0272727 = 0.00118078 loss) | |
| I0506 01:04:49.924249 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0331715 (* 0.0272727 = 0.000904676 loss) | |
| I0506 01:04:49.924263 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0287202 (* 0.0272727 = 0.000783279 loss) | |
| I0506 01:04:49.924278 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0123468 (* 0.0272727 = 0.000336731 loss) | |
| I0506 01:04:49.924291 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0110719 (* 0.0272727 = 0.00030196 loss) | |
| I0506 01:04:49.924305 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00888394 (* 0.0272727 = 0.000242289 loss) | |
| I0506 01:04:49.924319 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00831943 (* 0.0272727 = 0.000226894 loss) | |
| I0506 01:04:49.924331 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.104167 | |
| I0506 01:04:49.924343 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:04:49.924355 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:04:49.924367 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:04:49.924379 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 01:04:49.924391 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 01:04:49.924402 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 01:04:49.924414 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 01:04:49.924427 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:04:49.924437 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:04:49.924449 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:04:49.924460 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:04:49.924473 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:04:49.924484 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 01:04:49.924495 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 01:04:49.924506 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0506 01:04:49.924518 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:04:49.924530 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:04:49.924542 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:04:49.924551 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:04:49.924564 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:04:49.924576 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:04:49.924587 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:04:49.924599 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0506 01:04:49.924617 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0506 01:04:49.924633 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.30082 (* 0.3 = 0.990246 loss) | |
| I0506 01:04:49.924646 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.00934 (* 0.3 = 0.302803 loss) | |
| I0506 01:04:49.924664 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.25958 (* 0.0272727 = 0.0888977 loss) | |
| I0506 01:04:49.924679 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.15247 (* 0.0272727 = 0.0859763 loss) | |
| I0506 01:04:49.924705 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.74091 (* 0.0272727 = 0.102025 loss) | |
| I0506 01:04:49.924720 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 2.82792 (* 0.0272727 = 0.0771252 loss) | |
| I0506 01:04:49.924733 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.36506 (* 0.0272727 = 0.0645017 loss) | |
| I0506 01:04:49.924747 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.78531 (* 0.0272727 = 0.0486902 loss) | |
| I0506 01:04:49.924762 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.11509 (* 0.0272727 = 0.0304115 loss) | |
| I0506 01:04:49.924774 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.832551 (* 0.0272727 = 0.0227059 loss) | |
| I0506 01:04:49.924788 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.565274 (* 0.0272727 = 0.0154166 loss) | |
| I0506 01:04:49.924803 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.501374 (* 0.0272727 = 0.0136738 loss) | |
| I0506 01:04:49.924816 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.65265 (* 0.0272727 = 0.0177995 loss) | |
| I0506 01:04:49.924829 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.713095 (* 0.0272727 = 0.0194481 loss) | |
| I0506 01:04:49.924844 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.58373 (* 0.0272727 = 0.0159199 loss) | |
| I0506 01:04:49.924856 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.657769 (* 0.0272727 = 0.0179392 loss) | |
| I0506 01:04:49.924870 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.671718 (* 0.0272727 = 0.0183196 loss) | |
| I0506 01:04:49.924885 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0279034 (* 0.0272727 = 0.000761001 loss) | |
| I0506 01:04:49.924898 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0188801 (* 0.0272727 = 0.000514911 loss) | |
| I0506 01:04:49.924911 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0198359 (* 0.0272727 = 0.000540978 loss) | |
| I0506 01:04:49.924928 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.010802 (* 0.0272727 = 0.000294601 loss) | |
| I0506 01:04:49.924943 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00715719 (* 0.0272727 = 0.000195196 loss) | |
| I0506 01:04:49.924957 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00549443 (* 0.0272727 = 0.000149848 loss) | |
| I0506 01:04:49.924970 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00781407 (* 0.0272727 = 0.000213111 loss) | |
| I0506 01:04:49.924983 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0625 | |
| I0506 01:04:49.924994 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:04:49.925006 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0506 01:04:49.925017 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:04:49.925029 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:04:49.925040 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 01:04:49.925052 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 01:04:49.925063 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 01:04:49.925076 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 01:04:49.925086 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:04:49.925098 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:04:49.925109 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:04:49.925138 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:04:49.925153 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 01:04:49.925164 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 01:04:49.925176 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0506 01:04:49.925189 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:04:49.925211 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:04:49.925225 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:04:49.925236 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:04:49.925247 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:04:49.925258 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:04:49.925271 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:04:49.925282 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.738636 | |
| I0506 01:04:49.925293 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.270833 | |
| I0506 01:04:49.925307 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.12362 (* 1 = 3.12362 loss) | |
| I0506 01:04:49.925321 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.903029 (* 1 = 0.903029 loss) | |
| I0506 01:04:49.925335 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.14302 (* 0.0909091 = 0.285729 loss) | |
| I0506 01:04:49.925348 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.80875 (* 0.0909091 = 0.255341 loss) | |
| I0506 01:04:49.925362 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.31556 (* 0.0909091 = 0.301415 loss) | |
| I0506 01:04:49.925375 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.54021 (* 0.0909091 = 0.230928 loss) | |
| I0506 01:04:49.925390 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.14028 (* 0.0909091 = 0.194571 loss) | |
| I0506 01:04:49.925403 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.66334 (* 0.0909091 = 0.151213 loss) | |
| I0506 01:04:49.925416 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.00642 (* 0.0909091 = 0.0914927 loss) | |
| I0506 01:04:49.925431 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.609401 (* 0.0909091 = 0.0554001 loss) | |
| I0506 01:04:49.925444 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.431853 (* 0.0909091 = 0.0392594 loss) | |
| I0506 01:04:49.925457 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.361581 (* 0.0909091 = 0.032871 loss) | |
| I0506 01:04:49.925472 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.669643 (* 0.0909091 = 0.0608767 loss) | |
| I0506 01:04:49.925485 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.676325 (* 0.0909091 = 0.0614841 loss) | |
| I0506 01:04:49.925498 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.549119 (* 0.0909091 = 0.0499199 loss) | |
| I0506 01:04:49.925513 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.412904 (* 0.0909091 = 0.0375367 loss) | |
| I0506 01:04:49.925525 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.392425 (* 0.0909091 = 0.035675 loss) | |
| I0506 01:04:49.925539 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0735802 (* 0.0909091 = 0.00668911 loss) | |
| I0506 01:04:49.925554 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0612064 (* 0.0909091 = 0.00556422 loss) | |
| I0506 01:04:49.925567 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.029667 (* 0.0909091 = 0.002697 loss) | |
| I0506 01:04:49.925581 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0114395 (* 0.0909091 = 0.00103995 loss) | |
| I0506 01:04:49.925595 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00967834 (* 0.0909091 = 0.000879849 loss) | |
| I0506 01:04:49.925608 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00606467 (* 0.0909091 = 0.000551333 loss) | |
| I0506 01:04:49.925622 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00399857 (* 0.0909091 = 0.000363507 loss) | |
| I0506 01:04:49.925634 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:04:49.925645 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:04:49.925657 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000490988 | |
| I0506 01:04:49.925678 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000741576 | |
| I0506 01:04:49.925693 15760 sgd_solver.cpp:106] Iteration 11500, lr = 0.001 | |
| I0506 01:05:21.972265 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 47.5154 > 30) by scale factor 0.631375 | |
| I0506 01:06:30.842111 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.7563 > 30) by scale factor 0.816187 | |
| I0506 01:06:35.330862 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0768 > 30) by scale factor 0.997446 | |
| I0506 01:06:37.308562 15760 solver.cpp:229] Iteration 12000, loss = 10.3015 | |
| I0506 01:06:37.308627 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0625 | |
| I0506 01:06:37.308646 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.375 | |
| I0506 01:06:37.308660 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:06:37.308671 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:06:37.308682 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 01:06:37.308694 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 01:06:37.308706 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 01:06:37.308718 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:06:37.308730 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 01:06:37.308743 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:06:37.308753 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:06:37.308765 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:06:37.308776 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:06:37.308789 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:06:37.308800 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:06:37.308818 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:06:37.308830 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:06:37.308843 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:06:37.308854 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:06:37.308866 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:06:37.308877 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:06:37.308889 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:06:37.308908 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:06:37.308920 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0506 01:06:37.308933 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.208333 | |
| I0506 01:06:37.308948 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.33432 (* 0.3 = 1.0003 loss) | |
| I0506 01:06:37.308962 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.05465 (* 0.3 = 0.316395 loss) | |
| I0506 01:06:37.308976 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.40606 (* 0.0272727 = 0.0928927 loss) | |
| I0506 01:06:37.308990 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.32203 (* 0.0272727 = 0.0906007 loss) | |
| I0506 01:06:37.309005 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.76209 (* 0.0272727 = 0.102603 loss) | |
| I0506 01:06:37.309018 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.9733 (* 0.0272727 = 0.108363 loss) | |
| I0506 01:06:37.309031 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.51999 (* 0.0272727 = 0.0959996 loss) | |
| I0506 01:06:37.309046 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.55716 (* 0.0272727 = 0.0970133 loss) | |
| I0506 01:06:37.309058 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.35234 (* 0.0272727 = 0.0641547 loss) | |
| I0506 01:06:37.309072 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.28462 (* 0.0272727 = 0.00776238 loss) | |
| I0506 01:06:37.309087 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.212122 (* 0.0272727 = 0.00578514 loss) | |
| I0506 01:06:37.309101 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.100715 (* 0.0272727 = 0.00274677 loss) | |
| I0506 01:06:37.309114 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.104065 (* 0.0272727 = 0.00283815 loss) | |
| I0506 01:06:37.309175 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0633845 (* 0.0272727 = 0.00172867 loss) | |
| I0506 01:06:37.309191 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.058751 (* 0.0272727 = 0.0016023 loss) | |
| I0506 01:06:37.309206 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0433146 (* 0.0272727 = 0.00118131 loss) | |
| I0506 01:06:37.309219 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0222198 (* 0.0272727 = 0.000605995 loss) | |
| I0506 01:06:37.309233 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0179303 (* 0.0272727 = 0.000489008 loss) | |
| I0506 01:06:37.309247 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00814499 (* 0.0272727 = 0.000222136 loss) | |
| I0506 01:06:37.309262 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0112573 (* 0.0272727 = 0.000307016 loss) | |
| I0506 01:06:37.309275 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00863001 (* 0.0272727 = 0.000235364 loss) | |
| I0506 01:06:37.309289 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00559227 (* 0.0272727 = 0.000152516 loss) | |
| I0506 01:06:37.309303 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0061799 (* 0.0272727 = 0.000168543 loss) | |
| I0506 01:06:37.309317 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00416933 (* 0.0272727 = 0.000113709 loss) | |
| I0506 01:06:37.309329 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.125 | |
| I0506 01:06:37.309340 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:06:37.309352 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:06:37.309365 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:06:37.309376 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:06:37.309387 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 01:06:37.309396 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:06:37.309403 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:06:37.309415 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 01:06:37.309427 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:06:37.309438 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:06:37.309449 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:06:37.309468 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:06:37.309479 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:06:37.309491 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:06:37.309502 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:06:37.309514 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:06:37.309525 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:06:37.309535 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:06:37.309546 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:06:37.309558 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:06:37.309569 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:06:37.309581 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:06:37.309592 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0506 01:06:37.309603 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.291667 | |
| I0506 01:06:37.309617 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.265 (* 0.3 = 0.979499 loss) | |
| I0506 01:06:37.309631 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.978868 (* 0.3 = 0.29366 loss) | |
| I0506 01:06:37.309645 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.6756 (* 0.0272727 = 0.100244 loss) | |
| I0506 01:06:37.309671 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.40544 (* 0.0272727 = 0.0928756 loss) | |
| I0506 01:06:37.309686 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.65614 (* 0.0272727 = 0.0997128 loss) | |
| I0506 01:06:37.309700 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.57142 (* 0.0272727 = 0.0974024 loss) | |
| I0506 01:06:37.309713 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.12843 (* 0.0272727 = 0.0853207 loss) | |
| I0506 01:06:37.309727 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.20829 (* 0.0272727 = 0.0874988 loss) | |
| I0506 01:06:37.309741 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.29678 (* 0.0272727 = 0.0626396 loss) | |
| I0506 01:06:37.309756 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.176046 (* 0.0272727 = 0.00480124 loss) | |
| I0506 01:06:37.309769 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.135367 (* 0.0272727 = 0.00369182 loss) | |
| I0506 01:06:37.309782 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0449097 (* 0.0272727 = 0.00122481 loss) | |
| I0506 01:06:37.309797 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0228787 (* 0.0272727 = 0.000623964 loss) | |
| I0506 01:06:37.309810 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0326295 (* 0.0272727 = 0.000889896 loss) | |
| I0506 01:06:37.309824 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0373717 (* 0.0272727 = 0.00101923 loss) | |
| I0506 01:06:37.309839 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0109434 (* 0.0272727 = 0.000298455 loss) | |
| I0506 01:06:37.309851 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0140702 (* 0.0272727 = 0.000383732 loss) | |
| I0506 01:06:37.309870 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0124313 (* 0.0272727 = 0.000339035 loss) | |
| I0506 01:06:37.309885 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00905197 (* 0.0272727 = 0.000246872 loss) | |
| I0506 01:06:37.309898 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00837656 (* 0.0272727 = 0.000228452 loss) | |
| I0506 01:06:37.309912 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00606906 (* 0.0272727 = 0.00016552 loss) | |
| I0506 01:06:37.309926 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00706285 (* 0.0272727 = 0.000192623 loss) | |
| I0506 01:06:37.309940 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00581072 (* 0.0272727 = 0.000158474 loss) | |
| I0506 01:06:37.309954 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00417125 (* 0.0272727 = 0.000113761 loss) | |
| I0506 01:06:37.309967 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0416667 | |
| I0506 01:06:37.309978 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:06:37.309989 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:06:37.310000 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:06:37.310012 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:06:37.310024 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 01:06:37.310034 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 01:06:37.310046 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:06:37.310057 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 01:06:37.310070 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:06:37.310081 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:06:37.310091 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:06:37.310102 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:06:37.310113 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:06:37.310125 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:06:37.310147 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:06:37.310159 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:06:37.310170 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:06:37.310183 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:06:37.310194 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:06:37.310205 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:06:37.310216 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:06:37.310227 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:06:37.310238 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0506 01:06:37.310250 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.229167 | |
| I0506 01:06:37.310263 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.98158 (* 1 = 2.98158 loss) | |
| I0506 01:06:37.310277 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.899339 (* 1 = 0.899339 loss) | |
| I0506 01:06:37.310291 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.07966 (* 0.0909091 = 0.279969 loss) | |
| I0506 01:06:37.310304 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.28932 (* 0.0909091 = 0.299029 loss) | |
| I0506 01:06:37.310317 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.15286 (* 0.0909091 = 0.286624 loss) | |
| I0506 01:06:37.310331 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.19431 (* 0.0909091 = 0.290392 loss) | |
| I0506 01:06:37.310344 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.68118 (* 0.0909091 = 0.243744 loss) | |
| I0506 01:06:37.310358 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.52793 (* 0.0909091 = 0.229812 loss) | |
| I0506 01:06:37.310371 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.7011 (* 0.0909091 = 0.154646 loss) | |
| I0506 01:06:37.310385 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.137951 (* 0.0909091 = 0.012541 loss) | |
| I0506 01:06:37.310398 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0577708 (* 0.0909091 = 0.00525189 loss) | |
| I0506 01:06:37.310412 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0237347 (* 0.0909091 = 0.0021577 loss) | |
| I0506 01:06:37.310426 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00681033 (* 0.0909091 = 0.000619121 loss) | |
| I0506 01:06:37.310441 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0037311 (* 0.0909091 = 0.000339191 loss) | |
| I0506 01:06:37.310454 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00319353 (* 0.0909091 = 0.000290321 loss) | |
| I0506 01:06:37.310467 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00437012 (* 0.0909091 = 0.000397284 loss) | |
| I0506 01:06:37.310482 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00475531 (* 0.0909091 = 0.000432301 loss) | |
| I0506 01:06:37.310495 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00240352 (* 0.0909091 = 0.000218502 loss) | |
| I0506 01:06:37.310509 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00163277 (* 0.0909091 = 0.000148434 loss) | |
| I0506 01:06:37.310523 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00154774 (* 0.0909091 = 0.000140704 loss) | |
| I0506 01:06:37.310536 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000717736 (* 0.0909091 = 6.52487e-05 loss) | |
| I0506 01:06:37.310550 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000817442 (* 0.0909091 = 7.43129e-05 loss) | |
| I0506 01:06:37.310564 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000807638 (* 0.0909091 = 7.34217e-05 loss) | |
| I0506 01:06:37.310578 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0007146 (* 0.0909091 = 6.49637e-05 loss) | |
| I0506 01:06:37.310590 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:06:37.310611 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:06:37.310622 15760 solver.cpp:245] Train net output #149: total_confidence = 2.16203e-05 | |
| I0506 01:06:37.310634 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 9.1891e-05 | |
| I0506 01:06:37.310647 15760 sgd_solver.cpp:106] Iteration 12000, lr = 0.001 | |
| I0506 01:06:44.426525 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.5373 > 30) by scale factor 0.922018 | |
| I0506 01:06:55.568876 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.9125 > 30) by scale factor 0.791296 | |
| I0506 01:07:31.619340 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.129 > 30) by scale factor 0.995718 | |
| I0506 01:07:34.632398 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.5592 > 30) by scale factor 0.893942 | |
| I0506 01:08:01.907382 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.0591 > 30) by scale factor 0.965901 | |
| I0506 01:08:24.633149 15760 solver.cpp:229] Iteration 12500, loss = 10.3307 | |
| I0506 01:08:24.633213 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.08 | |
| I0506 01:08:24.633230 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:08:24.633244 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 01:08:24.633255 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:08:24.633267 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:08:24.633280 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:08:24.633291 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:08:24.633302 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:08:24.633316 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 01:08:24.633327 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 01:08:24.633340 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:08:24.633352 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:08:24.633363 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:08:24.633374 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:08:24.633386 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:08:24.633397 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:08:24.633409 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:08:24.633421 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:08:24.633432 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:08:24.633445 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:08:24.633465 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:08:24.633476 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:08:24.633488 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:08:24.633499 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.721591 | |
| I0506 01:08:24.633512 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.24 | |
| I0506 01:08:24.633527 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.28076 (* 0.3 = 0.98423 loss) | |
| I0506 01:08:24.633543 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.16126 (* 0.3 = 0.348378 loss) | |
| I0506 01:08:24.633555 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.38701 (* 0.0272727 = 0.0923729 loss) | |
| I0506 01:08:24.633569 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.76501 (* 0.0272727 = 0.102682 loss) | |
| I0506 01:08:24.633584 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 4.18069 (* 0.0272727 = 0.114019 loss) | |
| I0506 01:08:24.633597 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.25799 (* 0.0272727 = 0.0888544 loss) | |
| I0506 01:08:24.633611 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.67981 (* 0.0272727 = 0.0730857 loss) | |
| I0506 01:08:24.633625 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.59701 (* 0.0272727 = 0.0708276 loss) | |
| I0506 01:08:24.633638 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.33479 (* 0.0272727 = 0.0636761 loss) | |
| I0506 01:08:24.633652 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 2.14751 (* 0.0272727 = 0.0585684 loss) | |
| I0506 01:08:24.633666 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.45953 (* 0.0272727 = 0.0398055 loss) | |
| I0506 01:08:24.633679 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.216417 (* 0.0272727 = 0.00590229 loss) | |
| I0506 01:08:24.633693 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.162257 (* 0.0272727 = 0.00442518 loss) | |
| I0506 01:08:24.633708 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0965052 (* 0.0272727 = 0.00263196 loss) | |
| I0506 01:08:24.633752 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0801705 (* 0.0272727 = 0.00218647 loss) | |
| I0506 01:08:24.633769 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0657624 (* 0.0272727 = 0.00179352 loss) | |
| I0506 01:08:24.633782 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0375755 (* 0.0272727 = 0.00102479 loss) | |
| I0506 01:08:24.633796 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0179099 (* 0.0272727 = 0.000488451 loss) | |
| I0506 01:08:24.633817 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0221905 (* 0.0272727 = 0.000605196 loss) | |
| I0506 01:08:24.633832 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00476622 (* 0.0272727 = 0.000129988 loss) | |
| I0506 01:08:24.633846 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0102008 (* 0.0272727 = 0.000278205 loss) | |
| I0506 01:08:24.633860 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00531479 (* 0.0272727 = 0.000144949 loss) | |
| I0506 01:08:24.633873 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00434108 (* 0.0272727 = 0.000118393 loss) | |
| I0506 01:08:24.633891 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0032883 (* 0.0272727 = 8.96808e-05 loss) | |
| I0506 01:08:24.633904 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.02 | |
| I0506 01:08:24.633918 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:08:24.633929 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0506 01:08:24.633940 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:08:24.633952 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:08:24.633963 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:08:24.633975 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:08:24.633986 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:08:24.633997 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:08:24.634009 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 01:08:24.634030 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:08:24.634050 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:08:24.634063 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:08:24.634075 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:08:24.634086 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:08:24.634098 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:08:24.634109 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:08:24.634120 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:08:24.634131 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:08:24.634142 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:08:24.634155 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:08:24.634166 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:08:24.634176 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:08:24.634187 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.681818 | |
| I0506 01:08:24.634199 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.16 | |
| I0506 01:08:24.634212 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.32224 (* 0.3 = 0.996672 loss) | |
| I0506 01:08:24.634227 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.25783 (* 0.3 = 0.377348 loss) | |
| I0506 01:08:24.634240 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.47466 (* 0.0272727 = 0.0947634 loss) | |
| I0506 01:08:24.634254 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.07151 (* 0.0272727 = 0.0837685 loss) | |
| I0506 01:08:24.634280 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.57248 (* 0.0272727 = 0.0974312 loss) | |
| I0506 01:08:24.634295 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.6985 (* 0.0272727 = 0.100868 loss) | |
| I0506 01:08:24.634308 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.50879 (* 0.0272727 = 0.0684215 loss) | |
| I0506 01:08:24.634322 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.41572 (* 0.0272727 = 0.0658834 loss) | |
| I0506 01:08:24.634336 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.02353 (* 0.0272727 = 0.0551871 loss) | |
| I0506 01:08:24.634349 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.67112 (* 0.0272727 = 0.0455759 loss) | |
| I0506 01:08:24.634363 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.39346 (* 0.0272727 = 0.0380034 loss) | |
| I0506 01:08:24.634377 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.397091 (* 0.0272727 = 0.0108298 loss) | |
| I0506 01:08:24.634392 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.179715 (* 0.0272727 = 0.00490131 loss) | |
| I0506 01:08:24.634405 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.222558 (* 0.0272727 = 0.00606976 loss) | |
| I0506 01:08:24.634418 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.140311 (* 0.0272727 = 0.00382665 loss) | |
| I0506 01:08:24.634433 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0848269 (* 0.0272727 = 0.00231346 loss) | |
| I0506 01:08:24.634446 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0706339 (* 0.0272727 = 0.00192638 loss) | |
| I0506 01:08:24.634460 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0258409 (* 0.0272727 = 0.000704751 loss) | |
| I0506 01:08:24.634474 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0131587 (* 0.0272727 = 0.000358874 loss) | |
| I0506 01:08:24.634487 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0137824 (* 0.0272727 = 0.000375885 loss) | |
| I0506 01:08:24.634501 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00651639 (* 0.0272727 = 0.00017772 loss) | |
| I0506 01:08:24.634516 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0104886 (* 0.0272727 = 0.000286054 loss) | |
| I0506 01:08:24.634528 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0131744 (* 0.0272727 = 0.000359301 loss) | |
| I0506 01:08:24.634542 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0105068 (* 0.0272727 = 0.00028655 loss) | |
| I0506 01:08:24.634554 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.1 | |
| I0506 01:08:24.634567 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:08:24.634577 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0506 01:08:24.634589 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:08:24.634600 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:08:24.634613 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:08:24.634624 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:08:24.634634 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:08:24.634645 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0506 01:08:24.634657 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 01:08:24.634668 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:08:24.634680 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:08:24.634690 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:08:24.634702 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:08:24.634713 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:08:24.634724 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:08:24.634735 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:08:24.634757 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:08:24.634769 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:08:24.634780 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:08:24.634791 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:08:24.634802 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:08:24.634814 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:08:24.634824 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.727273 | |
| I0506 01:08:24.634836 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.32 | |
| I0506 01:08:24.634850 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.11879 (* 1 = 3.11879 loss) | |
| I0506 01:08:24.634863 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.05873 (* 1 = 1.05873 loss) | |
| I0506 01:08:24.634873 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.99311 (* 0.0909091 = 0.272101 loss) | |
| I0506 01:08:24.634887 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.13939 (* 0.0909091 = 0.285399 loss) | |
| I0506 01:08:24.634902 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.08555 (* 0.0909091 = 0.280504 loss) | |
| I0506 01:08:24.634914 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.79411 (* 0.0909091 = 0.25401 loss) | |
| I0506 01:08:24.634927 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.76 (* 0.0909091 = 0.250909 loss) | |
| I0506 01:08:24.634946 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.31418 (* 0.0909091 = 0.21038 loss) | |
| I0506 01:08:24.634968 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.67582 (* 0.0909091 = 0.152347 loss) | |
| I0506 01:08:24.634987 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.40122 (* 0.0909091 = 0.127384 loss) | |
| I0506 01:08:24.635001 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.3908 (* 0.0909091 = 0.126436 loss) | |
| I0506 01:08:24.635015 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.254715 (* 0.0909091 = 0.0231559 loss) | |
| I0506 01:08:24.635028 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.178407 (* 0.0909091 = 0.0162188 loss) | |
| I0506 01:08:24.635042 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0836328 (* 0.0909091 = 0.00760298 loss) | |
| I0506 01:08:24.635056 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0828859 (* 0.0909091 = 0.00753509 loss) | |
| I0506 01:08:24.635069 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0505152 (* 0.0909091 = 0.00459229 loss) | |
| I0506 01:08:24.635083 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0262974 (* 0.0909091 = 0.00239068 loss) | |
| I0506 01:08:24.635097 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0246523 (* 0.0909091 = 0.00224112 loss) | |
| I0506 01:08:24.635110 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0128676 (* 0.0909091 = 0.00116979 loss) | |
| I0506 01:08:24.635124 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00776033 (* 0.0909091 = 0.000705485 loss) | |
| I0506 01:08:24.635138 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0036484 (* 0.0909091 = 0.000331673 loss) | |
| I0506 01:08:24.635151 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0014328 (* 0.0909091 = 0.000130254 loss) | |
| I0506 01:08:24.635165 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000516041 (* 0.0909091 = 4.69128e-05 loss) | |
| I0506 01:08:24.635179 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000200503 (* 0.0909091 = 1.82275e-05 loss) | |
| I0506 01:08:24.635191 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:08:24.635202 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:08:24.635213 15760 solver.cpp:245] Train net output #149: total_confidence = 8.77779e-05 | |
| I0506 01:08:24.635234 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000276919 | |
| I0506 01:08:24.635249 15760 sgd_solver.cpp:106] Iteration 12500, lr = 0.001 | |
| I0506 01:08:47.420627 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.5332 > 30) by scale factor 0.894634 | |
| I0506 01:09:36.976449 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.38 > 30) by scale factor 0.872602 | |
| I0506 01:09:49.866714 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 53.2879 > 30) by scale factor 0.562979 | |
| I0506 01:10:04.725905 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.0608 > 30) by scale factor 0.997977 | |
| I0506 01:10:12.517460 15760 solver.cpp:229] Iteration 13000, loss = 10.2157 | |
| I0506 01:10:12.517675 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0357143 | |
| I0506 01:10:12.517712 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:10:12.517731 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0506 01:10:12.517745 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:10:12.517757 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:10:12.517770 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0506 01:10:12.517788 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0506 01:10:12.517801 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:10:12.517812 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 01:10:12.517824 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.625 | |
| I0506 01:10:12.517837 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:10:12.517849 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:10:12.517860 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:10:12.517875 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 01:10:12.517889 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:10:12.517901 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:10:12.517913 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:10:12.517925 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:10:12.517936 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:10:12.517948 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:10:12.517961 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:10:12.517972 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:10:12.517984 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:10:12.517997 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.693182 | |
| I0506 01:10:12.518008 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.160714 | |
| I0506 01:10:12.518024 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.3727 (* 0.3 = 1.01181 loss) | |
| I0506 01:10:12.518038 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.14547 (* 0.3 = 0.34364 loss) | |
| I0506 01:10:12.518054 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.49963 (* 0.0272727 = 0.0954445 loss) | |
| I0506 01:10:12.518066 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.26925 (* 0.0272727 = 0.0891613 loss) | |
| I0506 01:10:12.518080 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.40375 (* 0.0272727 = 0.0928296 loss) | |
| I0506 01:10:12.518095 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.1232 (* 0.0272727 = 0.0851781 loss) | |
| I0506 01:10:12.518108 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.15812 (* 0.0272727 = 0.0588577 loss) | |
| I0506 01:10:12.518121 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.56126 (* 0.0272727 = 0.0698525 loss) | |
| I0506 01:10:12.518136 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.17122 (* 0.0272727 = 0.0592151 loss) | |
| I0506 01:10:12.518149 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.43488 (* 0.0272727 = 0.039133 loss) | |
| I0506 01:10:12.518163 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.56377 (* 0.0272727 = 0.0426484 loss) | |
| I0506 01:10:12.518177 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.568061 (* 0.0272727 = 0.0154926 loss) | |
| I0506 01:10:12.518192 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.777873 (* 0.0272727 = 0.0212147 loss) | |
| I0506 01:10:12.518204 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.523339 (* 0.0272727 = 0.0142729 loss) | |
| I0506 01:10:12.518218 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.64375 (* 0.0272727 = 0.0175568 loss) | |
| I0506 01:10:12.518400 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0912989 (* 0.0272727 = 0.00248997 loss) | |
| I0506 01:10:12.518419 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0253375 (* 0.0272727 = 0.000691024 loss) | |
| I0506 01:10:12.518442 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0224558 (* 0.0272727 = 0.000612432 loss) | |
| I0506 01:10:12.518476 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00506593 (* 0.0272727 = 0.000138162 loss) | |
| I0506 01:10:12.518498 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00373114 (* 0.0272727 = 0.000101758 loss) | |
| I0506 01:10:12.518514 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.011854 (* 0.0272727 = 0.000323291 loss) | |
| I0506 01:10:12.518528 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00297973 (* 0.0272727 = 8.12654e-05 loss) | |
| I0506 01:10:12.518543 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00365449 (* 0.0272727 = 9.96678e-05 loss) | |
| I0506 01:10:12.518556 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00317327 (* 0.0272727 = 8.65437e-05 loss) | |
| I0506 01:10:12.518568 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0506 01:10:12.518581 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:10:12.518592 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:10:12.518612 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 01:10:12.518625 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:10:12.518635 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 01:10:12.518647 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:10:12.518659 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:10:12.518671 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0506 01:10:12.518682 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.625 | |
| I0506 01:10:12.518694 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:10:12.518707 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:10:12.518718 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:10:12.518730 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 01:10:12.518743 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:10:12.518754 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:10:12.518765 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:10:12.518776 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:10:12.518789 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:10:12.518800 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:10:12.518811 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:10:12.518822 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:10:12.518833 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:10:12.518846 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.681818 | |
| I0506 01:10:12.518857 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.142857 | |
| I0506 01:10:12.518870 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.35831 (* 0.3 = 1.00749 loss) | |
| I0506 01:10:12.518888 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.16586 (* 0.3 = 0.349757 loss) | |
| I0506 01:10:12.518903 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.58595 (* 0.0272727 = 0.0977987 loss) | |
| I0506 01:10:12.518918 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.5601 (* 0.0272727 = 0.0970938 loss) | |
| I0506 01:10:12.518944 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.19653 (* 0.0272727 = 0.0871782 loss) | |
| I0506 01:10:12.518959 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 4.00132 (* 0.0272727 = 0.109127 loss) | |
| I0506 01:10:12.518973 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 1.95664 (* 0.0272727 = 0.0533629 loss) | |
| I0506 01:10:12.518996 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.583 (* 0.0272727 = 0.0704454 loss) | |
| I0506 01:10:12.519013 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.86355 (* 0.0272727 = 0.0508242 loss) | |
| I0506 01:10:12.519028 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.26932 (* 0.0272727 = 0.0346179 loss) | |
| I0506 01:10:12.519042 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.84542 (* 0.0272727 = 0.0503296 loss) | |
| I0506 01:10:12.519060 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.618624 (* 0.0272727 = 0.0168716 loss) | |
| I0506 01:10:12.519074 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.542232 (* 0.0272727 = 0.0147882 loss) | |
| I0506 01:10:12.519088 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.458524 (* 0.0272727 = 0.0125052 loss) | |
| I0506 01:10:12.519101 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.72497 (* 0.0272727 = 0.0197719 loss) | |
| I0506 01:10:12.519115 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0564237 (* 0.0272727 = 0.00153883 loss) | |
| I0506 01:10:12.519129 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0100467 (* 0.0272727 = 0.000274001 loss) | |
| I0506 01:10:12.519143 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0124112 (* 0.0272727 = 0.000338487 loss) | |
| I0506 01:10:12.519157 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00286671 (* 0.0272727 = 7.81829e-05 loss) | |
| I0506 01:10:12.519171 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00168184 (* 0.0272727 = 4.58683e-05 loss) | |
| I0506 01:10:12.519186 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00150142 (* 0.0272727 = 4.09478e-05 loss) | |
| I0506 01:10:12.519201 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00118066 (* 0.0272727 = 3.21998e-05 loss) | |
| I0506 01:10:12.519214 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.000862421 (* 0.0272727 = 2.35206e-05 loss) | |
| I0506 01:10:12.519229 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00118918 (* 0.0272727 = 3.24322e-05 loss) | |
| I0506 01:10:12.519242 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0714286 | |
| I0506 01:10:12.519253 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:10:12.519265 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:10:12.519278 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:10:12.519289 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:10:12.519301 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 01:10:12.519314 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:10:12.519325 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:10:12.519336 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0506 01:10:12.519348 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0506 01:10:12.519361 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:10:12.519371 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:10:12.519383 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:10:12.519395 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 01:10:12.519410 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:10:12.519423 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:10:12.519435 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:10:12.519457 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:10:12.519470 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:10:12.519482 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:10:12.519490 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:10:12.519498 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:10:12.519506 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:10:12.519520 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0506 01:10:12.519532 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.214286 | |
| I0506 01:10:12.519546 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.21378 (* 1 = 3.21378 loss) | |
| I0506 01:10:12.519567 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.11266 (* 1 = 1.11266 loss) | |
| I0506 01:10:12.519579 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.99447 (* 0.0909091 = 0.272224 loss) | |
| I0506 01:10:12.519593 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.26806 (* 0.0909091 = 0.297097 loss) | |
| I0506 01:10:12.519606 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.90881 (* 0.0909091 = 0.264437 loss) | |
| I0506 01:10:12.519620 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.15341 (* 0.0909091 = 0.286673 loss) | |
| I0506 01:10:12.519640 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.76352 (* 0.0909091 = 0.16032 loss) | |
| I0506 01:10:12.519654 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.28891 (* 0.0909091 = 0.208082 loss) | |
| I0506 01:10:12.519667 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.82664 (* 0.0909091 = 0.166058 loss) | |
| I0506 01:10:12.519680 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.31262 (* 0.0909091 = 0.119329 loss) | |
| I0506 01:10:12.519695 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.75423 (* 0.0909091 = 0.159475 loss) | |
| I0506 01:10:12.519707 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.520176 (* 0.0909091 = 0.0472887 loss) | |
| I0506 01:10:12.519721 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.558078 (* 0.0909091 = 0.0507344 loss) | |
| I0506 01:10:12.519734 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.538366 (* 0.0909091 = 0.0489423 loss) | |
| I0506 01:10:12.519748 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.625333 (* 0.0909091 = 0.0568485 loss) | |
| I0506 01:10:12.519762 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0624547 (* 0.0909091 = 0.0056777 loss) | |
| I0506 01:10:12.519775 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0194882 (* 0.0909091 = 0.00177165 loss) | |
| I0506 01:10:12.519789 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00930948 (* 0.0909091 = 0.000846316 loss) | |
| I0506 01:10:12.519804 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00805821 (* 0.0909091 = 0.000732565 loss) | |
| I0506 01:10:12.519816 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00322805 (* 0.0909091 = 0.000293459 loss) | |
| I0506 01:10:12.519830 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0023985 (* 0.0909091 = 0.000218045 loss) | |
| I0506 01:10:12.519843 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00179379 (* 0.0909091 = 0.000163072 loss) | |
| I0506 01:10:12.519857 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000596851 (* 0.0909091 = 5.42592e-05 loss) | |
| I0506 01:10:12.519871 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000275066 (* 0.0909091 = 2.5006e-05 loss) | |
| I0506 01:10:12.519883 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:10:12.519894 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:10:12.519906 15760 solver.cpp:245] Train net output #149: total_confidence = 2.13964e-05 | |
| I0506 01:10:12.519930 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 8.1941e-05 | |
| I0506 01:10:12.519947 15760 sgd_solver.cpp:106] Iteration 13000, lr = 0.001 | |
| I0506 01:10:52.888744 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.1151 > 30) by scale factor 0.808296 | |
| I0506 01:11:14.626556 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 56.7327 > 30) by scale factor 0.528795 | |
| I0506 01:11:41.191826 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.2491 > 30) by scale factor 0.93026 | |
| I0506 01:12:00.263945 15760 solver.cpp:229] Iteration 13500, loss = 10.1872 | |
| I0506 01:12:00.264014 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.128205 | |
| I0506 01:12:00.264030 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:12:00.264047 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 01:12:00.264060 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:12:00.264072 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:12:00.264084 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0506 01:12:00.264096 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 01:12:00.264108 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 01:12:00.264119 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 01:12:00.264132 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:12:00.264142 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:12:00.264154 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:12:00.264166 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:12:00.264178 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:12:00.264189 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:12:00.264200 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:12:00.264212 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:12:00.264225 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:12:00.264235 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:12:00.264247 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:12:00.264258 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:12:00.264271 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:12:00.264281 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:12:00.264293 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.806818 | |
| I0506 01:12:00.264304 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.205128 | |
| I0506 01:12:00.264322 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.29026 (* 0.3 = 0.987078 loss) | |
| I0506 01:12:00.264335 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.8184 (* 0.3 = 0.24552 loss) | |
| I0506 01:12:00.264358 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.03156 (* 0.0272727 = 0.0826789 loss) | |
| I0506 01:12:00.264374 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.61096 (* 0.0272727 = 0.0984808 loss) | |
| I0506 01:12:00.264387 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.62234 (* 0.0272727 = 0.098791 loss) | |
| I0506 01:12:00.264400 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.55539 (* 0.0272727 = 0.0969651 loss) | |
| I0506 01:12:00.264415 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.20658 (* 0.0272727 = 0.0601795 loss) | |
| I0506 01:12:00.264428 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.99262 (* 0.0272727 = 0.0543443 loss) | |
| I0506 01:12:00.264442 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.15218 (* 0.0272727 = 0.031423 loss) | |
| I0506 01:12:00.264456 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.148992 (* 0.0272727 = 0.00406341 loss) | |
| I0506 01:12:00.264470 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0792642 (* 0.0272727 = 0.00216175 loss) | |
| I0506 01:12:00.264484 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0480067 (* 0.0272727 = 0.00130927 loss) | |
| I0506 01:12:00.264498 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0326268 (* 0.0272727 = 0.000889821 loss) | |
| I0506 01:12:00.264513 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0384926 (* 0.0272727 = 0.0010498 loss) | |
| I0506 01:12:00.264556 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0275469 (* 0.0272727 = 0.000751279 loss) | |
| I0506 01:12:00.264571 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0288686 (* 0.0272727 = 0.000787325 loss) | |
| I0506 01:12:00.264585 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0203367 (* 0.0272727 = 0.000554636 loss) | |
| I0506 01:12:00.264600 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0205238 (* 0.0272727 = 0.000559741 loss) | |
| I0506 01:12:00.264613 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0120362 (* 0.0272727 = 0.00032826 loss) | |
| I0506 01:12:00.264627 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0120819 (* 0.0272727 = 0.000329505 loss) | |
| I0506 01:12:00.264642 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00940583 (* 0.0272727 = 0.000256523 loss) | |
| I0506 01:12:00.264655 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.012492 (* 0.0272727 = 0.000340691 loss) | |
| I0506 01:12:00.264669 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0130575 (* 0.0272727 = 0.000356113 loss) | |
| I0506 01:12:00.264683 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00789107 (* 0.0272727 = 0.000215211 loss) | |
| I0506 01:12:00.264695 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0512821 | |
| I0506 01:12:00.264708 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0506 01:12:00.264719 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:12:00.264730 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:12:00.264742 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 01:12:00.264753 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 01:12:00.264765 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 01:12:00.264775 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 01:12:00.264787 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 01:12:00.264798 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:12:00.264809 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:12:00.264821 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:12:00.264832 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:12:00.264842 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:12:00.264858 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:12:00.264869 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:12:00.264881 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:12:00.264892 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:12:00.264900 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:12:00.264907 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:12:00.264919 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:12:00.264930 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:12:00.264942 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:12:00.264953 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0506 01:12:00.264971 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.179487 | |
| I0506 01:12:00.264986 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.23658 (* 0.3 = 0.970975 loss) | |
| I0506 01:12:00.264999 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.810961 (* 0.3 = 0.243288 loss) | |
| I0506 01:12:00.265013 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.26388 (* 0.0272727 = 0.089015 loss) | |
| I0506 01:12:00.265027 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.65537 (* 0.0272727 = 0.0996919 loss) | |
| I0506 01:12:00.265053 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.69845 (* 0.0272727 = 0.100867 loss) | |
| I0506 01:12:00.265067 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.55433 (* 0.0272727 = 0.0969362 loss) | |
| I0506 01:12:00.265080 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.22202 (* 0.0272727 = 0.0606005 loss) | |
| I0506 01:12:00.265115 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.76985 (* 0.0272727 = 0.0482687 loss) | |
| I0506 01:12:00.265131 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.975666 (* 0.0272727 = 0.0266091 loss) | |
| I0506 01:12:00.265146 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.128465 (* 0.0272727 = 0.00350358 loss) | |
| I0506 01:12:00.265161 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0378701 (* 0.0272727 = 0.00103282 loss) | |
| I0506 01:12:00.265174 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0261739 (* 0.0272727 = 0.000713833 loss) | |
| I0506 01:12:00.265188 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0140477 (* 0.0272727 = 0.000383118 loss) | |
| I0506 01:12:00.265202 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0129944 (* 0.0272727 = 0.000354393 loss) | |
| I0506 01:12:00.265215 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.015936 (* 0.0272727 = 0.000434618 loss) | |
| I0506 01:12:00.265229 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00958437 (* 0.0272727 = 0.000261392 loss) | |
| I0506 01:12:00.265244 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0186364 (* 0.0272727 = 0.000508266 loss) | |
| I0506 01:12:00.265256 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0117355 (* 0.0272727 = 0.000320058 loss) | |
| I0506 01:12:00.265270 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.012361 (* 0.0272727 = 0.000337119 loss) | |
| I0506 01:12:00.265283 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0389757 (* 0.0272727 = 0.00106297 loss) | |
| I0506 01:12:00.265297 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00643344 (* 0.0272727 = 0.000175458 loss) | |
| I0506 01:12:00.265311 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0334483 (* 0.0272727 = 0.000912227 loss) | |
| I0506 01:12:00.265324 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0109932 (* 0.0272727 = 0.000299814 loss) | |
| I0506 01:12:00.265337 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0248216 (* 0.0272727 = 0.000676953 loss) | |
| I0506 01:12:00.265349 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.102564 | |
| I0506 01:12:00.265360 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 01:12:00.265372 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:12:00.265383 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:12:00.265395 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:12:00.265406 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0506 01:12:00.265418 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.75 | |
| I0506 01:12:00.265429 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 01:12:00.265439 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 01:12:00.265450 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:12:00.265462 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:12:00.265473 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:12:00.265485 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:12:00.265496 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:12:00.265506 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:12:00.265517 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:12:00.265540 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:12:00.265552 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:12:00.265564 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:12:00.265575 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:12:00.265586 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:12:00.265597 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:12:00.265609 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:12:00.265619 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.795455 | |
| I0506 01:12:00.265630 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.230769 | |
| I0506 01:12:00.265645 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.99007 (* 1 = 2.99007 loss) | |
| I0506 01:12:00.265658 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.74796 (* 1 = 0.74796 loss) | |
| I0506 01:12:00.265672 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.75477 (* 0.0909091 = 0.250434 loss) | |
| I0506 01:12:00.265686 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.07956 (* 0.0909091 = 0.27996 loss) | |
| I0506 01:12:00.265698 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.36443 (* 0.0909091 = 0.305857 loss) | |
| I0506 01:12:00.265712 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.34752 (* 0.0909091 = 0.30432 loss) | |
| I0506 01:12:00.265725 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.66298 (* 0.0909091 = 0.15118 loss) | |
| I0506 01:12:00.265739 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.2915 (* 0.0909091 = 0.117409 loss) | |
| I0506 01:12:00.265753 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.904418 (* 0.0909091 = 0.0822199 loss) | |
| I0506 01:12:00.265765 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.180925 (* 0.0909091 = 0.0164478 loss) | |
| I0506 01:12:00.265779 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0179006 (* 0.0909091 = 0.00162733 loss) | |
| I0506 01:12:00.265794 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0132754 (* 0.0909091 = 0.00120686 loss) | |
| I0506 01:12:00.265807 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00942598 (* 0.0909091 = 0.000856907 loss) | |
| I0506 01:12:00.265820 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00812456 (* 0.0909091 = 0.000738596 loss) | |
| I0506 01:12:00.265835 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00845734 (* 0.0909091 = 0.000768849 loss) | |
| I0506 01:12:00.265847 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00978049 (* 0.0909091 = 0.000889136 loss) | |
| I0506 01:12:00.265861 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00507534 (* 0.0909091 = 0.000461394 loss) | |
| I0506 01:12:00.265875 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00725744 (* 0.0909091 = 0.000659768 loss) | |
| I0506 01:12:00.265889 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00443012 (* 0.0909091 = 0.000402738 loss) | |
| I0506 01:12:00.265905 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00384406 (* 0.0909091 = 0.00034946 loss) | |
| I0506 01:12:00.265920 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00329056 (* 0.0909091 = 0.000299142 loss) | |
| I0506 01:12:00.265934 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00335404 (* 0.0909091 = 0.000304913 loss) | |
| I0506 01:12:00.265947 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00272079 (* 0.0909091 = 0.000247344 loss) | |
| I0506 01:12:00.265961 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00186497 (* 0.0909091 = 0.000169543 loss) | |
| I0506 01:12:00.265974 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:12:00.265985 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:12:00.266005 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000459506 | |
| I0506 01:12:00.266018 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00110744 | |
| I0506 01:12:00.266032 15760 sgd_solver.cpp:106] Iteration 13500, lr = 0.001 | |
| I0506 01:13:25.868455 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.6551 > 30) by scale factor 0.776094 | |
| I0506 01:13:47.521690 15760 solver.cpp:229] Iteration 14000, loss = 10.1379 | |
| I0506 01:13:47.521754 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.130435 | |
| I0506 01:13:47.521773 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:13:47.521786 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:13:47.521798 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.375 | |
| I0506 01:13:47.521809 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:13:47.521821 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:13:47.521833 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0506 01:13:47.521847 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:13:47.521858 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:13:47.521872 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:13:47.521886 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:13:47.521898 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:13:47.521910 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:13:47.521922 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:13:47.521936 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:13:47.521950 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:13:47.521960 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:13:47.521972 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:13:47.521984 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:13:47.521996 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:13:47.522007 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:13:47.522018 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:13:47.522030 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:13:47.522042 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.761364 | |
| I0506 01:13:47.522053 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.369565 | |
| I0506 01:13:47.522068 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.97803 (* 0.3 = 0.89341 loss) | |
| I0506 01:13:47.522083 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.00445 (* 0.3 = 0.301334 loss) | |
| I0506 01:13:47.522097 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.3012 (* 0.0272727 = 0.0900328 loss) | |
| I0506 01:13:47.522111 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 2.86598 (* 0.0272727 = 0.0781632 loss) | |
| I0506 01:13:47.522125 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.07876 (* 0.0272727 = 0.0839661 loss) | |
| I0506 01:13:47.522140 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.64763 (* 0.0272727 = 0.0722081 loss) | |
| I0506 01:13:47.522153 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.30138 (* 0.0272727 = 0.0627648 loss) | |
| I0506 01:13:47.522166 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.93026 (* 0.0272727 = 0.0526436 loss) | |
| I0506 01:13:47.522181 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.36061 (* 0.0272727 = 0.0371076 loss) | |
| I0506 01:13:47.522193 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.754575 (* 0.0272727 = 0.0205793 loss) | |
| I0506 01:13:47.522207 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.743939 (* 0.0272727 = 0.0202892 loss) | |
| I0506 01:13:47.522222 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.63626 (* 0.0272727 = 0.0173525 loss) | |
| I0506 01:13:47.522235 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0957874 (* 0.0272727 = 0.00261238 loss) | |
| I0506 01:13:47.522249 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0661391 (* 0.0272727 = 0.00180379 loss) | |
| I0506 01:13:47.522295 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0477949 (* 0.0272727 = 0.0013035 loss) | |
| I0506 01:13:47.522311 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.03091 (* 0.0272727 = 0.000843001 loss) | |
| I0506 01:13:47.522325 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0221597 (* 0.0272727 = 0.000604356 loss) | |
| I0506 01:13:47.522341 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.014267 (* 0.0272727 = 0.0003891 loss) | |
| I0506 01:13:47.522354 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0115984 (* 0.0272727 = 0.000316319 loss) | |
| I0506 01:13:47.522367 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0122472 (* 0.0272727 = 0.000334014 loss) | |
| I0506 01:13:47.522382 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00506716 (* 0.0272727 = 0.000138195 loss) | |
| I0506 01:13:47.522395 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00842909 (* 0.0272727 = 0.000229884 loss) | |
| I0506 01:13:47.522408 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00839747 (* 0.0272727 = 0.000229022 loss) | |
| I0506 01:13:47.522423 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0065175 (* 0.0272727 = 0.00017775 loss) | |
| I0506 01:13:47.522434 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.108696 | |
| I0506 01:13:47.522446 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0506 01:13:47.522459 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.375 | |
| I0506 01:13:47.522470 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:13:47.522482 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 01:13:47.522493 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:13:47.522505 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0506 01:13:47.522517 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:13:47.522528 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:13:47.522541 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:13:47.522552 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:13:47.522563 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:13:47.522575 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:13:47.522586 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:13:47.522598 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:13:47.522609 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:13:47.522620 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:13:47.522631 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:13:47.522644 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:13:47.522655 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:13:47.522666 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:13:47.522677 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:13:47.522689 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:13:47.522701 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0506 01:13:47.522711 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.347826 | |
| I0506 01:13:47.522725 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.98626 (* 0.3 = 0.895877 loss) | |
| I0506 01:13:47.522739 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.92827 (* 0.3 = 0.278481 loss) | |
| I0506 01:13:47.522753 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.18989 (* 0.0272727 = 0.0869969 loss) | |
| I0506 01:13:47.522766 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.7356 (* 0.0272727 = 0.0746073 loss) | |
| I0506 01:13:47.522790 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.66572 (* 0.0272727 = 0.0999741 loss) | |
| I0506 01:13:47.522805 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 2.85147 (* 0.0272727 = 0.0777674 loss) | |
| I0506 01:13:47.522819 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.36399 (* 0.0272727 = 0.0644726 loss) | |
| I0506 01:13:47.522832 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.24724 (* 0.0272727 = 0.0612882 loss) | |
| I0506 01:13:47.522846 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.88441 (* 0.0272727 = 0.0513929 loss) | |
| I0506 01:13:47.522860 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.928529 (* 0.0272727 = 0.0253235 loss) | |
| I0506 01:13:47.522873 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.767638 (* 0.0272727 = 0.0209356 loss) | |
| I0506 01:13:47.522886 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.52933 (* 0.0272727 = 0.0144363 loss) | |
| I0506 01:13:47.522897 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0556491 (* 0.0272727 = 0.0015177 loss) | |
| I0506 01:13:47.522913 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0656989 (* 0.0272727 = 0.00179179 loss) | |
| I0506 01:13:47.522930 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0627824 (* 0.0272727 = 0.00171225 loss) | |
| I0506 01:13:47.522945 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0198789 (* 0.0272727 = 0.000542153 loss) | |
| I0506 01:13:47.522959 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0284077 (* 0.0272727 = 0.000774756 loss) | |
| I0506 01:13:47.522974 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0101408 (* 0.0272727 = 0.000276566 loss) | |
| I0506 01:13:47.522990 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0160885 (* 0.0272727 = 0.000438776 loss) | |
| I0506 01:13:47.523005 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00902091 (* 0.0272727 = 0.000246025 loss) | |
| I0506 01:13:47.523018 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00520205 (* 0.0272727 = 0.000141874 loss) | |
| I0506 01:13:47.523032 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00560155 (* 0.0272727 = 0.00015277 loss) | |
| I0506 01:13:47.523046 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00532253 (* 0.0272727 = 0.00014516 loss) | |
| I0506 01:13:47.523059 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00504309 (* 0.0272727 = 0.000137539 loss) | |
| I0506 01:13:47.523072 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0869565 | |
| I0506 01:13:47.523083 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 01:13:47.523094 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 01:13:47.523105 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:13:47.523118 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:13:47.523128 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 01:13:47.523139 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 01:13:47.523150 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:13:47.523161 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:13:47.523174 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:13:47.523185 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:13:47.523195 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:13:47.523206 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:13:47.523217 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:13:47.523228 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:13:47.523239 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:13:47.523262 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:13:47.523273 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:13:47.523285 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:13:47.523296 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:13:47.523308 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:13:47.523319 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:13:47.523329 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:13:47.523340 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0506 01:13:47.523352 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.391304 | |
| I0506 01:13:47.523365 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.75523 (* 1 = 2.75523 loss) | |
| I0506 01:13:47.523380 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.808721 (* 1 = 0.808721 loss) | |
| I0506 01:13:47.523392 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.71116 (* 0.0909091 = 0.246469 loss) | |
| I0506 01:13:47.523406 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.39447 (* 0.0909091 = 0.21768 loss) | |
| I0506 01:13:47.523419 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.07068 (* 0.0909091 = 0.279153 loss) | |
| I0506 01:13:47.523433 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.23634 (* 0.0909091 = 0.203304 loss) | |
| I0506 01:13:47.523447 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.84821 (* 0.0909091 = 0.168019 loss) | |
| I0506 01:13:47.523459 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.67811 (* 0.0909091 = 0.152556 loss) | |
| I0506 01:13:47.523473 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.23269 (* 0.0909091 = 0.112063 loss) | |
| I0506 01:13:47.523486 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.729383 (* 0.0909091 = 0.0663075 loss) | |
| I0506 01:13:47.523500 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.556753 (* 0.0909091 = 0.0506139 loss) | |
| I0506 01:13:47.523514 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.473569 (* 0.0909091 = 0.0430517 loss) | |
| I0506 01:13:47.523526 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.168734 (* 0.0909091 = 0.0153394 loss) | |
| I0506 01:13:47.523540 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.13414 (* 0.0909091 = 0.0121946 loss) | |
| I0506 01:13:47.523553 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0574514 (* 0.0909091 = 0.00522286 loss) | |
| I0506 01:13:47.523567 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0399175 (* 0.0909091 = 0.00362886 loss) | |
| I0506 01:13:47.523581 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0205918 (* 0.0909091 = 0.00187198 loss) | |
| I0506 01:13:47.523594 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0129964 (* 0.0909091 = 0.00118149 loss) | |
| I0506 01:13:47.523608 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00545606 (* 0.0909091 = 0.000496006 loss) | |
| I0506 01:13:47.523622 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00573461 (* 0.0909091 = 0.000521328 loss) | |
| I0506 01:13:47.523635 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00281748 (* 0.0909091 = 0.000256134 loss) | |
| I0506 01:13:47.523649 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000925993 (* 0.0909091 = 8.41812e-05 loss) | |
| I0506 01:13:47.523663 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00137911 (* 0.0909091 = 0.000125373 loss) | |
| I0506 01:13:47.523676 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000760699 (* 0.0909091 = 6.91545e-05 loss) | |
| I0506 01:13:47.523689 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:13:47.523700 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:13:47.523720 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000513099 | |
| I0506 01:13:47.523733 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00227681 | |
| I0506 01:13:47.523747 15760 sgd_solver.cpp:106] Iteration 14000, lr = 0.001 | |
| I0506 01:13:54.205555 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.205 > 30) by scale factor 0.993213 | |
| I0506 01:14:10.259616 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 53.6161 > 30) by scale factor 0.559533 | |
| I0506 01:14:11.775970 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.7052 > 30) by scale factor 0.775089 | |
| I0506 01:15:34.706714 15760 solver.cpp:229] Iteration 14500, loss = 10.1471 | |
| I0506 01:15:34.706827 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0909091 | |
| I0506 01:15:34.706850 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:15:34.706861 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:15:34.706876 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:15:34.706889 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 01:15:34.706902 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0506 01:15:34.706912 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0506 01:15:34.706924 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:15:34.706936 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:15:34.706948 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:15:34.706959 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:15:34.706971 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:15:34.706982 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:15:34.706993 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:15:34.707005 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:15:34.707015 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:15:34.707027 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:15:34.707038 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:15:34.707051 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:15:34.707062 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:15:34.707072 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:15:34.707084 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:15:34.707095 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:15:34.707106 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0506 01:15:34.707118 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.327273 | |
| I0506 01:15:34.707134 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.13687 (* 0.3 = 0.941061 loss) | |
| I0506 01:15:34.707147 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.05194 (* 0.3 = 0.315583 loss) | |
| I0506 01:15:34.707161 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.82961 (* 0.0272727 = 0.0771712 loss) | |
| I0506 01:15:34.707175 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.22008 (* 0.0272727 = 0.0878203 loss) | |
| I0506 01:15:34.707190 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.76604 (* 0.0272727 = 0.10271 loss) | |
| I0506 01:15:34.707202 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.68135 (* 0.0272727 = 0.1004 loss) | |
| I0506 01:15:34.707216 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.86663 (* 0.0272727 = 0.105454 loss) | |
| I0506 01:15:34.707231 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.22887 (* 0.0272727 = 0.0880601 loss) | |
| I0506 01:15:34.707244 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.09255 (* 0.0272727 = 0.0570697 loss) | |
| I0506 01:15:34.707258 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.83195 (* 0.0272727 = 0.0499622 loss) | |
| I0506 01:15:34.707272 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.651819 (* 0.0272727 = 0.0177769 loss) | |
| I0506 01:15:34.707285 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.844457 (* 0.0272727 = 0.0230306 loss) | |
| I0506 01:15:34.707299 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0500671 (* 0.0272727 = 0.00136547 loss) | |
| I0506 01:15:34.707314 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0688767 (* 0.0272727 = 0.00187846 loss) | |
| I0506 01:15:34.707327 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0371768 (* 0.0272727 = 0.00101391 loss) | |
| I0506 01:15:34.707360 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0269354 (* 0.0272727 = 0.000734603 loss) | |
| I0506 01:15:34.707375 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0297297 (* 0.0272727 = 0.00081081 loss) | |
| I0506 01:15:34.707389 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0213514 (* 0.0272727 = 0.00058231 loss) | |
| I0506 01:15:34.707402 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0176239 (* 0.0272727 = 0.000480651 loss) | |
| I0506 01:15:34.707417 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0103552 (* 0.0272727 = 0.000282414 loss) | |
| I0506 01:15:34.707430 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.01404 (* 0.0272727 = 0.000382909 loss) | |
| I0506 01:15:34.707444 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00743621 (* 0.0272727 = 0.000202806 loss) | |
| I0506 01:15:34.707458 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0115278 (* 0.0272727 = 0.000314394 loss) | |
| I0506 01:15:34.707473 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00793353 (* 0.0272727 = 0.000216369 loss) | |
| I0506 01:15:34.707484 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0909091 | |
| I0506 01:15:34.707496 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0506 01:15:34.707509 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:15:34.707520 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:15:34.707531 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 01:15:34.707542 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 01:15:34.707553 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0506 01:15:34.707566 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:15:34.707576 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:15:34.707587 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:15:34.707599 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:15:34.707610 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:15:34.707622 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:15:34.707633 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:15:34.707643 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:15:34.707654 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:15:34.707666 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:15:34.707677 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:15:34.707689 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:15:34.707700 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:15:34.707710 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:15:34.707721 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:15:34.707732 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:15:34.707743 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.715909 | |
| I0506 01:15:34.707754 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.272727 | |
| I0506 01:15:34.707768 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.1986 (* 0.3 = 0.959581 loss) | |
| I0506 01:15:34.707782 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.06007 (* 0.3 = 0.318021 loss) | |
| I0506 01:15:34.707797 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.31494 (* 0.0272727 = 0.0904075 loss) | |
| I0506 01:15:34.707810 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.26341 (* 0.0272727 = 0.0890022 loss) | |
| I0506 01:15:34.707839 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.45696 (* 0.0272727 = 0.0942807 loss) | |
| I0506 01:15:34.707872 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.48927 (* 0.0272727 = 0.095162 loss) | |
| I0506 01:15:34.707890 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.51656 (* 0.0272727 = 0.0959062 loss) | |
| I0506 01:15:34.707906 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.29819 (* 0.0272727 = 0.0899505 loss) | |
| I0506 01:15:34.707918 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.12806 (* 0.0272727 = 0.058038 loss) | |
| I0506 01:15:34.707936 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.6017 (* 0.0272727 = 0.0436827 loss) | |
| I0506 01:15:34.707949 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.460615 (* 0.0272727 = 0.0125622 loss) | |
| I0506 01:15:34.707963 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.704828 (* 0.0272727 = 0.0192226 loss) | |
| I0506 01:15:34.707978 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0693191 (* 0.0272727 = 0.00189052 loss) | |
| I0506 01:15:34.707991 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0439532 (* 0.0272727 = 0.00119872 loss) | |
| I0506 01:15:34.708005 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0354576 (* 0.0272727 = 0.000967025 loss) | |
| I0506 01:15:34.708019 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0217756 (* 0.0272727 = 0.00059388 loss) | |
| I0506 01:15:34.708032 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0215318 (* 0.0272727 = 0.000587231 loss) | |
| I0506 01:15:34.708046 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0191593 (* 0.0272727 = 0.000522525 loss) | |
| I0506 01:15:34.708060 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00968549 (* 0.0272727 = 0.00026415 loss) | |
| I0506 01:15:34.708073 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0156878 (* 0.0272727 = 0.000427848 loss) | |
| I0506 01:15:34.708087 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00863561 (* 0.0272727 = 0.000235517 loss) | |
| I0506 01:15:34.708097 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00538772 (* 0.0272727 = 0.000146938 loss) | |
| I0506 01:15:34.708106 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0043156 (* 0.0272727 = 0.000117698 loss) | |
| I0506 01:15:34.708122 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00816693 (* 0.0272727 = 0.000222735 loss) | |
| I0506 01:15:34.708133 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.109091 | |
| I0506 01:15:34.708145 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:15:34.708158 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:15:34.708168 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:15:34.708180 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 01:15:34.708191 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0506 01:15:34.708202 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 01:15:34.708214 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:15:34.708225 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:15:34.708236 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:15:34.708247 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:15:34.708259 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:15:34.708271 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:15:34.708281 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:15:34.708292 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:15:34.708303 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:15:34.708314 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:15:34.708336 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:15:34.708349 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:15:34.708360 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:15:34.708371 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:15:34.708382 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:15:34.708395 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:15:34.708405 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.715909 | |
| I0506 01:15:34.708416 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.218182 | |
| I0506 01:15:34.708431 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.95003 (* 1 = 2.95003 loss) | |
| I0506 01:15:34.708443 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.969615 (* 1 = 0.969615 loss) | |
| I0506 01:15:34.708457 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.77837 (* 0.0909091 = 0.252579 loss) | |
| I0506 01:15:34.708470 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.00608 (* 0.0909091 = 0.27328 loss) | |
| I0506 01:15:34.708484 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.28518 (* 0.0909091 = 0.298653 loss) | |
| I0506 01:15:34.708498 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.34226 (* 0.0909091 = 0.303841 loss) | |
| I0506 01:15:34.708511 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.26479 (* 0.0909091 = 0.296799 loss) | |
| I0506 01:15:34.708524 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.90251 (* 0.0909091 = 0.263864 loss) | |
| I0506 01:15:34.708537 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.66449 (* 0.0909091 = 0.151317 loss) | |
| I0506 01:15:34.708550 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.13884 (* 0.0909091 = 0.103531 loss) | |
| I0506 01:15:34.708564 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.453446 (* 0.0909091 = 0.0412224 loss) | |
| I0506 01:15:34.708577 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.538978 (* 0.0909091 = 0.048998 loss) | |
| I0506 01:15:34.708591 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00492277 (* 0.0909091 = 0.000447525 loss) | |
| I0506 01:15:34.708605 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0100114 (* 0.0909091 = 0.00091013 loss) | |
| I0506 01:15:34.708618 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00538368 (* 0.0909091 = 0.000489426 loss) | |
| I0506 01:15:34.708631 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00470833 (* 0.0909091 = 0.00042803 loss) | |
| I0506 01:15:34.708645 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00311807 (* 0.0909091 = 0.000283461 loss) | |
| I0506 01:15:34.708659 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00493102 (* 0.0909091 = 0.000448274 loss) | |
| I0506 01:15:34.708673 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00168771 (* 0.0909091 = 0.000153428 loss) | |
| I0506 01:15:34.708686 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00183872 (* 0.0909091 = 0.000167156 loss) | |
| I0506 01:15:34.708700 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000734499 (* 0.0909091 = 6.67727e-05 loss) | |
| I0506 01:15:34.708714 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00067965 (* 0.0909091 = 6.17864e-05 loss) | |
| I0506 01:15:34.708729 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000349837 (* 0.0909091 = 3.18034e-05 loss) | |
| I0506 01:15:34.708741 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000503578 (* 0.0909091 = 4.57799e-05 loss) | |
| I0506 01:15:34.708753 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:15:34.708765 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:15:34.708776 15760 solver.cpp:245] Train net output #149: total_confidence = 3.93226e-07 | |
| I0506 01:15:34.708796 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.91091e-06 | |
| I0506 01:15:34.708811 15760 sgd_solver.cpp:106] Iteration 14500, lr = 0.001 | |
| I0506 01:17:21.871377 15760 solver.cpp:338] Iteration 15000, Testing net (#0) | |
| I0506 01:17:58.382797 15760 solver.cpp:393] Test loss: 9.46736 | |
| I0506 01:17:58.382943 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.064083 | |
| I0506 01:17:58.382964 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.117 | |
| I0506 01:17:58.382978 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.092 | |
| I0506 01:17:58.382990 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.072 | |
| I0506 01:17:58.383003 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.181 | |
| I0506 01:17:58.383015 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.292 | |
| I0506 01:17:58.383026 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.445 | |
| I0506 01:17:58.383038 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.727 | |
| I0506 01:17:58.383049 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.908 | |
| I0506 01:17:58.383060 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.99 | |
| I0506 01:17:58.383071 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0506 01:17:58.383082 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0506 01:17:58.383095 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0506 01:17:58.383105 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0506 01:17:58.383116 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0506 01:17:58.383128 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0506 01:17:58.383139 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0506 01:17:58.383152 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0506 01:17:58.383162 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0506 01:17:58.383173 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0506 01:17:58.383184 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0506 01:17:58.383195 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 01:17:58.383206 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 01:17:58.383218 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.763638 | |
| I0506 01:17:58.383229 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.227932 | |
| I0506 01:17:58.383245 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.74962 (* 0.3 = 1.12489 loss) | |
| I0506 01:17:58.383260 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.977147 (* 0.3 = 0.293144 loss) | |
| I0506 01:17:58.383273 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.15864 (* 0.0272727 = 0.0861447 loss) | |
| I0506 01:17:58.383286 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.29323 (* 0.0272727 = 0.0898153 loss) | |
| I0506 01:17:58.383299 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.40568 (* 0.0272727 = 0.0928823 loss) | |
| I0506 01:17:58.383312 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.20364 (* 0.0272727 = 0.0873721 loss) | |
| I0506 01:17:58.383327 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 2.81908 (* 0.0272727 = 0.0768841 loss) | |
| I0506 01:17:58.383339 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.38377 (* 0.0272727 = 0.0650119 loss) | |
| I0506 01:17:58.383352 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.40049 (* 0.0272727 = 0.0381953 loss) | |
| I0506 01:17:58.383365 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 0.539722 (* 0.0272727 = 0.0147197 loss) | |
| I0506 01:17:58.383379 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.0798035 (* 0.0272727 = 0.00217646 loss) | |
| I0506 01:17:58.383393 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0302285 (* 0.0272727 = 0.000824414 loss) | |
| I0506 01:17:58.383406 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.020605 (* 0.0272727 = 0.000561954 loss) | |
| I0506 01:17:58.383420 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0147296 (* 0.0272727 = 0.000401717 loss) | |
| I0506 01:17:58.383433 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0120589 (* 0.0272727 = 0.000328879 loss) | |
| I0506 01:17:58.383467 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.00989585 (* 0.0272727 = 0.000269887 loss) | |
| I0506 01:17:58.383482 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.00806179 (* 0.0272727 = 0.000219867 loss) | |
| I0506 01:17:58.383496 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.0057349 (* 0.0272727 = 0.000156406 loss) | |
| I0506 01:17:58.383509 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.00459388 (* 0.0272727 = 0.000125288 loss) | |
| I0506 01:17:58.383523 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00354053 (* 0.0272727 = 9.656e-05 loss) | |
| I0506 01:17:58.383538 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00299916 (* 0.0272727 = 8.17953e-05 loss) | |
| I0506 01:17:58.383551 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00346931 (* 0.0272727 = 9.46177e-05 loss) | |
| I0506 01:17:58.383565 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00316964 (* 0.0272727 = 8.64448e-05 loss) | |
| I0506 01:17:58.383579 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00307601 (* 0.0272727 = 8.38911e-05 loss) | |
| I0506 01:17:58.383590 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0610802 | |
| I0506 01:17:58.383602 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.109 | |
| I0506 01:17:58.383613 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.096 | |
| I0506 01:17:58.383625 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.079 | |
| I0506 01:17:58.383636 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.172 | |
| I0506 01:17:58.383646 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.289 | |
| I0506 01:17:58.383657 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.444 | |
| I0506 01:17:58.383669 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.727 | |
| I0506 01:17:58.383680 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.908 | |
| I0506 01:17:58.383692 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.99 | |
| I0506 01:17:58.383702 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0506 01:17:58.383713 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0506 01:17:58.383723 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0506 01:17:58.383734 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0506 01:17:58.383745 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0506 01:17:58.383756 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0506 01:17:58.383767 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0506 01:17:58.383779 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0506 01:17:58.383786 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0506 01:17:58.383793 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0506 01:17:58.383805 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0506 01:17:58.383816 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 01:17:58.383826 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 01:17:58.383837 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.76282 | |
| I0506 01:17:58.383848 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.230079 | |
| I0506 01:17:58.383862 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.72993 (* 0.3 = 1.11898 loss) | |
| I0506 01:17:58.383878 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.967423 (* 0.3 = 0.290227 loss) | |
| I0506 01:17:58.383893 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 3.08892 (* 0.0272727 = 0.0842433 loss) | |
| I0506 01:17:58.383906 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.20182 (* 0.0272727 = 0.0873224 loss) | |
| I0506 01:17:58.383919 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.30703 (* 0.0272727 = 0.0901918 loss) | |
| I0506 01:17:58.383944 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.12759 (* 0.0272727 = 0.0852979 loss) | |
| I0506 01:17:58.383963 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 2.75173 (* 0.0272727 = 0.0750472 loss) | |
| I0506 01:17:58.383977 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.34594 (* 0.0272727 = 0.0639803 loss) | |
| I0506 01:17:58.383991 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.37992 (* 0.0272727 = 0.0376341 loss) | |
| I0506 01:17:58.384003 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 0.541194 (* 0.0272727 = 0.0147598 loss) | |
| I0506 01:17:58.384017 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.0707106 (* 0.0272727 = 0.00192847 loss) | |
| I0506 01:17:58.384030 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0256099 (* 0.0272727 = 0.000698452 loss) | |
| I0506 01:17:58.384044 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0151536 (* 0.0272727 = 0.000413281 loss) | |
| I0506 01:17:58.384057 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0113996 (* 0.0272727 = 0.000310898 loss) | |
| I0506 01:17:58.384070 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.00836933 (* 0.0272727 = 0.000228255 loss) | |
| I0506 01:17:58.384084 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.00724379 (* 0.0272727 = 0.000197558 loss) | |
| I0506 01:17:58.384098 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00558723 (* 0.0272727 = 0.000152379 loss) | |
| I0506 01:17:58.384111 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00342645 (* 0.0272727 = 9.34487e-05 loss) | |
| I0506 01:17:58.384124 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.00220963 (* 0.0272727 = 6.02627e-05 loss) | |
| I0506 01:17:58.384137 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00207886 (* 0.0272727 = 5.66961e-05 loss) | |
| I0506 01:17:58.384151 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0017325 (* 0.0272727 = 4.725e-05 loss) | |
| I0506 01:17:58.384165 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00173066 (* 0.0272727 = 4.71998e-05 loss) | |
| I0506 01:17:58.384178 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00148852 (* 0.0272727 = 4.05961e-05 loss) | |
| I0506 01:17:58.384191 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00165501 (* 0.0272727 = 4.51367e-05 loss) | |
| I0506 01:17:58.384203 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0857385 | |
| I0506 01:17:58.384214 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.136 | |
| I0506 01:17:58.384225 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.111 | |
| I0506 01:17:58.384238 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.093 | |
| I0506 01:17:58.384248 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.191 | |
| I0506 01:17:58.384258 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.316 | |
| I0506 01:17:58.384269 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.449 | |
| I0506 01:17:58.384281 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.728 | |
| I0506 01:17:58.384291 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.908 | |
| I0506 01:17:58.384304 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.99 | |
| I0506 01:17:58.384315 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.998 | |
| I0506 01:17:58.384325 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.999 | |
| I0506 01:17:58.384336 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0506 01:17:58.384347 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0506 01:17:58.384358 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0506 01:17:58.384369 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0506 01:17:58.384379 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0506 01:17:58.384400 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0506 01:17:58.384413 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0506 01:17:58.384424 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0506 01:17:58.384435 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0506 01:17:58.384446 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 01:17:58.384456 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 01:17:58.384467 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.766092 | |
| I0506 01:17:58.384479 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.256556 | |
| I0506 01:17:58.384491 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.06403 (* 1 = 3.06403 loss) | |
| I0506 01:17:58.384505 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.821588 (* 1 = 0.821588 loss) | |
| I0506 01:17:58.384517 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 2.82631 (* 0.0909091 = 0.256937 loss) | |
| I0506 01:17:58.384531 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 2.99119 (* 0.0909091 = 0.271926 loss) | |
| I0506 01:17:58.384543 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.10522 (* 0.0909091 = 0.282292 loss) | |
| I0506 01:17:58.384557 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 2.87108 (* 0.0909091 = 0.261007 loss) | |
| I0506 01:17:58.384569 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 2.47231 (* 0.0909091 = 0.224755 loss) | |
| I0506 01:17:58.384582 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.11961 (* 0.0909091 = 0.192692 loss) | |
| I0506 01:17:58.384595 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.1951 (* 0.0909091 = 0.108645 loss) | |
| I0506 01:17:58.384608 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.463354 (* 0.0909091 = 0.0421231 loss) | |
| I0506 01:17:58.384621 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.0706262 (* 0.0909091 = 0.00642056 loss) | |
| I0506 01:17:58.384635 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0261324 (* 0.0909091 = 0.00237567 loss) | |
| I0506 01:17:58.384649 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0156067 (* 0.0909091 = 0.00141879 loss) | |
| I0506 01:17:58.384661 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0110937 (* 0.0909091 = 0.00100852 loss) | |
| I0506 01:17:58.384675 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.00949553 (* 0.0909091 = 0.00086323 loss) | |
| I0506 01:17:58.384687 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.00829717 (* 0.0909091 = 0.000754288 loss) | |
| I0506 01:17:58.384701 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.00691062 (* 0.0909091 = 0.000628238 loss) | |
| I0506 01:17:58.384714 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00454091 (* 0.0909091 = 0.00041281 loss) | |
| I0506 01:17:58.384727 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00260643 (* 0.0909091 = 0.000236949 loss) | |
| I0506 01:17:58.384742 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00225509 (* 0.0909091 = 0.000205008 loss) | |
| I0506 01:17:58.384754 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00156607 (* 0.0909091 = 0.00014237 loss) | |
| I0506 01:17:58.384768 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00139005 (* 0.0909091 = 0.000126369 loss) | |
| I0506 01:17:58.384781 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00120476 (* 0.0909091 = 0.000109524 loss) | |
| I0506 01:17:58.384795 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00109171 (* 0.0909091 = 9.92465e-05 loss) | |
| I0506 01:17:58.384807 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 01:17:58.384819 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:17:58.384829 15760 solver.cpp:406] Test net output #149: total_confidence = 0.000688981 | |
| I0506 01:17:58.384840 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.00127372 | |
| I0506 01:17:58.384862 15760 solver.cpp:338] Iteration 15000, Testing net (#1) | |
| I0506 01:18:34.738880 15760 solver.cpp:393] Test loss: 10.0826 | |
| I0506 01:18:34.739002 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0642857 | |
| I0506 01:18:34.739022 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.104 | |
| I0506 01:18:34.739035 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.084 | |
| I0506 01:18:34.739048 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.1 | |
| I0506 01:18:34.739059 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.188 | |
| I0506 01:18:34.739070 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.296 | |
| I0506 01:18:34.739083 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.414 | |
| I0506 01:18:34.739094 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.638 | |
| I0506 01:18:34.739104 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.794 | |
| I0506 01:18:34.739115 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.887 | |
| I0506 01:18:34.739127 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.91 | |
| I0506 01:18:34.739138 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.924 | |
| I0506 01:18:34.739150 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.934 | |
| I0506 01:18:34.739161 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.948 | |
| I0506 01:18:34.739172 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.959 | |
| I0506 01:18:34.739183 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.97 | |
| I0506 01:18:34.739194 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.98 | |
| I0506 01:18:34.739207 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.99 | |
| I0506 01:18:34.739217 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0506 01:18:34.739229 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0506 01:18:34.739240 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.999 | |
| I0506 01:18:34.739251 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 01:18:34.739264 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 01:18:34.739274 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.730228 | |
| I0506 01:18:34.739286 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.225088 | |
| I0506 01:18:34.739301 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.72146 (* 0.3 = 1.11644 loss) | |
| I0506 01:18:34.739318 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.10789 (* 0.3 = 0.332367 loss) | |
| I0506 01:18:34.739341 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.13767 (* 0.0272727 = 0.0855729 loss) | |
| I0506 01:18:34.739357 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.26956 (* 0.0272727 = 0.0891698 loss) | |
| I0506 01:18:34.739370 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.34981 (* 0.0272727 = 0.0913585 loss) | |
| I0506 01:18:34.739384 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.18527 (* 0.0272727 = 0.0868711 loss) | |
| I0506 01:18:34.739398 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 2.8337 (* 0.0272727 = 0.0772828 loss) | |
| I0506 01:18:34.739410 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.44116 (* 0.0272727 = 0.0665772 loss) | |
| I0506 01:18:34.739423 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.64405 (* 0.0272727 = 0.0448377 loss) | |
| I0506 01:18:34.739436 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 0.95023 (* 0.0272727 = 0.0259154 loss) | |
| I0506 01:18:34.739449 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.49676 (* 0.0272727 = 0.013548 loss) | |
| I0506 01:18:34.739462 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.416388 (* 0.0272727 = 0.011356 loss) | |
| I0506 01:18:34.739476 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.36524 (* 0.0272727 = 0.00996108 loss) | |
| I0506 01:18:34.739490 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.339022 (* 0.0272727 = 0.00924604 loss) | |
| I0506 01:18:34.739503 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.279407 (* 0.0272727 = 0.00762018 loss) | |
| I0506 01:18:34.739537 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.23614 (* 0.0272727 = 0.00644018 loss) | |
| I0506 01:18:34.739552 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.178836 (* 0.0272727 = 0.00487736 loss) | |
| I0506 01:18:34.739564 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.13203 (* 0.0272727 = 0.00360082 loss) | |
| I0506 01:18:34.739578 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0786726 (* 0.0272727 = 0.00214562 loss) | |
| I0506 01:18:34.739591 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0365071 (* 0.0272727 = 0.000995647 loss) | |
| I0506 01:18:34.739604 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0288934 (* 0.0272727 = 0.000788002 loss) | |
| I0506 01:18:34.739619 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.011237 (* 0.0272727 = 0.000306463 loss) | |
| I0506 01:18:34.739631 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00260023 (* 0.0272727 = 7.09155e-05 loss) | |
| I0506 01:18:34.739645 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00254123 (* 0.0272727 = 6.93062e-05 loss) | |
| I0506 01:18:34.739657 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0593548 | |
| I0506 01:18:34.739668 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.112 | |
| I0506 01:18:34.739680 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.097 | |
| I0506 01:18:34.739691 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.085 | |
| I0506 01:18:34.739702 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.193 | |
| I0506 01:18:34.739713 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.289 | |
| I0506 01:18:34.739724 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.414 | |
| I0506 01:18:34.739735 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.638 | |
| I0506 01:18:34.739747 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.796 | |
| I0506 01:18:34.739758 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.887 | |
| I0506 01:18:34.739771 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.91 | |
| I0506 01:18:34.739784 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.924 | |
| I0506 01:18:34.739795 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.934 | |
| I0506 01:18:34.739806 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.948 | |
| I0506 01:18:34.739819 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.959 | |
| I0506 01:18:34.739830 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.97 | |
| I0506 01:18:34.739840 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.98 | |
| I0506 01:18:34.739852 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.99 | |
| I0506 01:18:34.739863 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0506 01:18:34.739874 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0506 01:18:34.739886 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.999 | |
| I0506 01:18:34.739897 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 01:18:34.739907 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 01:18:34.739918 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.72991 | |
| I0506 01:18:34.739933 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.231621 | |
| I0506 01:18:34.739946 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.71242 (* 0.3 = 1.11373 loss) | |
| I0506 01:18:34.739960 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.09671 (* 0.3 = 0.329012 loss) | |
| I0506 01:18:34.739977 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 3.07392 (* 0.0272727 = 0.0838341 loss) | |
| I0506 01:18:34.739987 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.19504 (* 0.0272727 = 0.0871375 loss) | |
| I0506 01:18:34.740018 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.26694 (* 0.0272727 = 0.0890983 loss) | |
| I0506 01:18:34.740037 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.11242 (* 0.0272727 = 0.0848842 loss) | |
| I0506 01:18:34.740051 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 2.78209 (* 0.0272727 = 0.0758753 loss) | |
| I0506 01:18:34.740063 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.40416 (* 0.0272727 = 0.0655679 loss) | |
| I0506 01:18:34.740077 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.63179 (* 0.0272727 = 0.0445034 loss) | |
| I0506 01:18:34.740089 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 0.945957 (* 0.0272727 = 0.0257988 loss) | |
| I0506 01:18:34.740103 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.488492 (* 0.0272727 = 0.0133225 loss) | |
| I0506 01:18:34.740115 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.40967 (* 0.0272727 = 0.0111728 loss) | |
| I0506 01:18:34.740129 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.364839 (* 0.0272727 = 0.00995015 loss) | |
| I0506 01:18:34.740142 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.337633 (* 0.0272727 = 0.00920816 loss) | |
| I0506 01:18:34.740155 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.277557 (* 0.0272727 = 0.00756975 loss) | |
| I0506 01:18:34.740170 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.237736 (* 0.0272727 = 0.00648371 loss) | |
| I0506 01:18:34.740182 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.179616 (* 0.0272727 = 0.00489863 loss) | |
| I0506 01:18:34.740195 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.13734 (* 0.0272727 = 0.00374563 loss) | |
| I0506 01:18:34.740208 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0826075 (* 0.0272727 = 0.00225293 loss) | |
| I0506 01:18:34.740221 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0336986 (* 0.0272727 = 0.000919052 loss) | |
| I0506 01:18:34.740236 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.027736 (* 0.0272727 = 0.000756436 loss) | |
| I0506 01:18:34.740248 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0102872 (* 0.0272727 = 0.000280561 loss) | |
| I0506 01:18:34.740262 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00170133 (* 0.0272727 = 4.64e-05 loss) | |
| I0506 01:18:34.740274 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00165171 (* 0.0272727 = 4.50467e-05 loss) | |
| I0506 01:18:34.740286 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.081825 | |
| I0506 01:18:34.740298 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.138 | |
| I0506 01:18:34.740309 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.115 | |
| I0506 01:18:34.740319 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.091 | |
| I0506 01:18:34.740330 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.194 | |
| I0506 01:18:34.740341 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.317 | |
| I0506 01:18:34.740352 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.441 | |
| I0506 01:18:34.740363 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.642 | |
| I0506 01:18:34.740375 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.799 | |
| I0506 01:18:34.740386 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.893 | |
| I0506 01:18:34.740396 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.913 | |
| I0506 01:18:34.740407 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.924 | |
| I0506 01:18:34.740418 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.934 | |
| I0506 01:18:34.740429 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.948 | |
| I0506 01:18:34.740440 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.959 | |
| I0506 01:18:34.740452 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.97 | |
| I0506 01:18:34.740461 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.98 | |
| I0506 01:18:34.740483 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.99 | |
| I0506 01:18:34.740495 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0506 01:18:34.740507 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0506 01:18:34.740519 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.999 | |
| I0506 01:18:34.740530 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 01:18:34.740540 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 01:18:34.740550 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.7345 | |
| I0506 01:18:34.740561 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.261215 | |
| I0506 01:18:34.740576 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.07178 (* 1 = 3.07178 loss) | |
| I0506 01:18:34.740588 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.940189 (* 1 = 0.940189 loss) | |
| I0506 01:18:34.740602 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 2.84173 (* 0.0909091 = 0.258339 loss) | |
| I0506 01:18:34.740614 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 2.99783 (* 0.0909091 = 0.27253 loss) | |
| I0506 01:18:34.740628 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.07917 (* 0.0909091 = 0.279925 loss) | |
| I0506 01:18:34.740640 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 2.91793 (* 0.0909091 = 0.265266 loss) | |
| I0506 01:18:34.740653 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 2.52294 (* 0.0909091 = 0.229359 loss) | |
| I0506 01:18:34.740666 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.16318 (* 0.0909091 = 0.196653 loss) | |
| I0506 01:18:34.740679 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.44195 (* 0.0909091 = 0.131087 loss) | |
| I0506 01:18:34.740692 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.850479 (* 0.0909091 = 0.0773163 loss) | |
| I0506 01:18:34.740706 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.439606 (* 0.0909091 = 0.0399641 loss) | |
| I0506 01:18:34.740720 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.357882 (* 0.0909091 = 0.0325348 loss) | |
| I0506 01:18:34.740732 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.311207 (* 0.0909091 = 0.0282915 loss) | |
| I0506 01:18:34.740746 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.289679 (* 0.0909091 = 0.0263345 loss) | |
| I0506 01:18:34.740758 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.232066 (* 0.0909091 = 0.0210969 loss) | |
| I0506 01:18:34.740772 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.195612 (* 0.0909091 = 0.0177829 loss) | |
| I0506 01:18:34.740785 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.149774 (* 0.0909091 = 0.0136158 loss) | |
| I0506 01:18:34.740798 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.115048 (* 0.0909091 = 0.0104589 loss) | |
| I0506 01:18:34.740811 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0694714 (* 0.0909091 = 0.00631558 loss) | |
| I0506 01:18:34.740829 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0333526 (* 0.0909091 = 0.00303205 loss) | |
| I0506 01:18:34.740844 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0256266 (* 0.0909091 = 0.00232969 loss) | |
| I0506 01:18:34.740856 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00869203 (* 0.0909091 = 0.000790184 loss) | |
| I0506 01:18:34.740870 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00104604 (* 0.0909091 = 9.50944e-05 loss) | |
| I0506 01:18:34.740885 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.000635208 (* 0.0909091 = 5.77462e-05 loss) | |
| I0506 01:18:34.740895 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 01:18:34.740906 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:18:34.740917 15760 solver.cpp:406] Test net output #149: total_confidence = 0.000799994 | |
| I0506 01:18:34.740938 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.00130653 | |
| I0506 01:18:34.877380 15760 solver.cpp:229] Iteration 15000, loss = 10.1085 | |
| I0506 01:18:34.877436 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.025 | |
| I0506 01:18:34.877454 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:18:34.877466 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:18:34.877478 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:18:34.877490 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:18:34.877506 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:18:34.877519 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0506 01:18:34.877532 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 1 | |
| I0506 01:18:34.877544 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 01:18:34.877557 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:18:34.877568 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:18:34.877580 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:18:34.877591 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:18:34.877604 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:18:34.877614 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:18:34.877626 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:18:34.877638 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:18:34.877650 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:18:34.877662 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:18:34.877673 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:18:34.877686 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:18:34.877696 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:18:34.877708 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:18:34.877719 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0506 01:18:34.877732 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.125 | |
| I0506 01:18:34.877748 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.71756 (* 0.3 = 1.11527 loss) | |
| I0506 01:18:34.877761 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.963816 (* 0.3 = 0.289145 loss) | |
| I0506 01:18:34.877775 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.87316 (* 0.0272727 = 0.105632 loss) | |
| I0506 01:18:34.877789 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.61535 (* 0.0272727 = 0.0986003 loss) | |
| I0506 01:18:34.877802 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.56355 (* 0.0272727 = 0.0971877 loss) | |
| I0506 01:18:34.877816 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.18504 (* 0.0272727 = 0.0868647 loss) | |
| I0506 01:18:34.877830 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.87338 (* 0.0272727 = 0.078365 loss) | |
| I0506 01:18:34.877843 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.53873 (* 0.0272727 = 0.069238 loss) | |
| I0506 01:18:34.877857 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.157052 (* 0.0272727 = 0.00428324 loss) | |
| I0506 01:18:34.877871 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0487337 (* 0.0272727 = 0.0013291 loss) | |
| I0506 01:18:34.877887 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0302051 (* 0.0272727 = 0.000823777 loss) | |
| I0506 01:18:34.877900 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0205137 (* 0.0272727 = 0.000559463 loss) | |
| I0506 01:18:34.877918 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0152702 (* 0.0272727 = 0.000416461 loss) | |
| I0506 01:18:34.877961 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0102163 (* 0.0272727 = 0.000278625 loss) | |
| I0506 01:18:34.877976 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00797908 (* 0.0272727 = 0.000217611 loss) | |
| I0506 01:18:34.877990 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00660342 (* 0.0272727 = 0.000180093 loss) | |
| I0506 01:18:34.878005 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00706615 (* 0.0272727 = 0.000192713 loss) | |
| I0506 01:18:34.878018 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0063529 (* 0.0272727 = 0.000173261 loss) | |
| I0506 01:18:34.878032 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00324002 (* 0.0272727 = 8.83641e-05 loss) | |
| I0506 01:18:34.878046 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00122857 (* 0.0272727 = 3.35066e-05 loss) | |
| I0506 01:18:34.878059 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00253774 (* 0.0272727 = 6.9211e-05 loss) | |
| I0506 01:18:34.878073 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00240224 (* 0.0272727 = 6.55157e-05 loss) | |
| I0506 01:18:34.878087 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.000803981 (* 0.0272727 = 2.19268e-05 loss) | |
| I0506 01:18:34.878100 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000907273 (* 0.0272727 = 2.47438e-05 loss) | |
| I0506 01:18:34.878113 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.025 | |
| I0506 01:18:34.878125 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:18:34.878136 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:18:34.878149 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:18:34.878160 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:18:34.878171 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:18:34.878183 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0506 01:18:34.878195 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 1 | |
| I0506 01:18:34.878206 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 01:18:34.878217 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:18:34.878228 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:18:34.878240 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:18:34.878252 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:18:34.878262 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:18:34.878273 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:18:34.878285 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:18:34.878296 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:18:34.878307 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:18:34.878319 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:18:34.878330 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:18:34.878341 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:18:34.878353 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:18:34.878365 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:18:34.878376 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0506 01:18:34.878387 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.1 | |
| I0506 01:18:34.878401 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.62474 (* 0.3 = 1.08742 loss) | |
| I0506 01:18:34.878414 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.88559 (* 0.3 = 0.265677 loss) | |
| I0506 01:18:34.878428 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.68353 (* 0.0272727 = 0.10046 loss) | |
| I0506 01:18:34.878453 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.5538 (* 0.0272727 = 0.0969219 loss) | |
| I0506 01:18:34.878466 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.82261 (* 0.0272727 = 0.104253 loss) | |
| I0506 01:18:34.878480 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.55625 (* 0.0272727 = 0.0969887 loss) | |
| I0506 01:18:34.878494 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.86324 (* 0.0272727 = 0.0780884 loss) | |
| I0506 01:18:34.878507 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.39896 (* 0.0272727 = 0.0654261 loss) | |
| I0506 01:18:34.878521 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.134045 (* 0.0272727 = 0.00365577 loss) | |
| I0506 01:18:34.878535 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0501041 (* 0.0272727 = 0.00136648 loss) | |
| I0506 01:18:34.878548 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00912019 (* 0.0272727 = 0.000248733 loss) | |
| I0506 01:18:34.878567 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0139905 (* 0.0272727 = 0.000381558 loss) | |
| I0506 01:18:34.878582 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00842633 (* 0.0272727 = 0.000229809 loss) | |
| I0506 01:18:34.878595 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0138354 (* 0.0272727 = 0.000377328 loss) | |
| I0506 01:18:34.878609 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00810878 (* 0.0272727 = 0.000221149 loss) | |
| I0506 01:18:34.878623 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00586413 (* 0.0272727 = 0.000159931 loss) | |
| I0506 01:18:34.878638 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00505185 (* 0.0272727 = 0.000137778 loss) | |
| I0506 01:18:34.878650 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00277365 (* 0.0272727 = 7.5645e-05 loss) | |
| I0506 01:18:34.878664 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00213047 (* 0.0272727 = 5.81038e-05 loss) | |
| I0506 01:18:34.878679 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0020606 (* 0.0272727 = 5.61982e-05 loss) | |
| I0506 01:18:34.878692 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00109202 (* 0.0272727 = 2.97823e-05 loss) | |
| I0506 01:18:34.878706 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.000672109 (* 0.0272727 = 1.83303e-05 loss) | |
| I0506 01:18:34.878720 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00118826 (* 0.0272727 = 3.24072e-05 loss) | |
| I0506 01:18:34.878733 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.000796421 (* 0.0272727 = 2.17206e-05 loss) | |
| I0506 01:18:34.878746 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.05 | |
| I0506 01:18:34.878756 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:18:34.878768 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 01:18:34.878779 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:18:34.878792 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:18:34.878803 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 01:18:34.878813 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 01:18:34.878825 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 1 | |
| I0506 01:18:34.878836 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 01:18:34.878847 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:18:34.878859 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:18:34.878870 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:18:34.878882 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:18:34.878893 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:18:34.878901 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:18:34.878918 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:18:34.878931 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:18:34.878943 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:18:34.878957 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:18:34.878969 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:18:34.878981 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:18:34.878993 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:18:34.879004 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:18:34.879014 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.784091 | |
| I0506 01:18:34.879026 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.175 | |
| I0506 01:18:34.879040 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.43985 (* 1 = 3.43985 loss) | |
| I0506 01:18:34.879053 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.824498 (* 1 = 0.824498 loss) | |
| I0506 01:18:34.879066 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.50474 (* 0.0909091 = 0.318613 loss) | |
| I0506 01:18:34.879081 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.466 (* 0.0909091 = 0.315091 loss) | |
| I0506 01:18:34.879093 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.84658 (* 0.0909091 = 0.349689 loss) | |
| I0506 01:18:34.879107 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.06476 (* 0.0909091 = 0.278615 loss) | |
| I0506 01:18:34.879120 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.71465 (* 0.0909091 = 0.246786 loss) | |
| I0506 01:18:34.879133 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.02498 (* 0.0909091 = 0.184089 loss) | |
| I0506 01:18:34.879148 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.0592063 (* 0.0909091 = 0.00538239 loss) | |
| I0506 01:18:34.879160 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0386987 (* 0.0909091 = 0.00351806 loss) | |
| I0506 01:18:34.879174 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0363173 (* 0.0909091 = 0.00330157 loss) | |
| I0506 01:18:34.879187 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0360351 (* 0.0909091 = 0.00327592 loss) | |
| I0506 01:18:34.879201 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0316278 (* 0.0909091 = 0.00287525 loss) | |
| I0506 01:18:34.879215 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0190793 (* 0.0909091 = 0.00173448 loss) | |
| I0506 01:18:34.879228 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.033115 (* 0.0909091 = 0.00301046 loss) | |
| I0506 01:18:34.879241 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0245007 (* 0.0909091 = 0.00222734 loss) | |
| I0506 01:18:34.879254 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0149691 (* 0.0909091 = 0.00136083 loss) | |
| I0506 01:18:34.879268 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0144147 (* 0.0909091 = 0.00131043 loss) | |
| I0506 01:18:34.879281 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00605087 (* 0.0909091 = 0.000550079 loss) | |
| I0506 01:18:34.879295 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00368396 (* 0.0909091 = 0.000334906 loss) | |
| I0506 01:18:34.879309 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00407637 (* 0.0909091 = 0.000370579 loss) | |
| I0506 01:18:34.879323 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00311162 (* 0.0909091 = 0.000282874 loss) | |
| I0506 01:18:34.879336 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00273891 (* 0.0909091 = 0.000248992 loss) | |
| I0506 01:18:34.879349 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00181369 (* 0.0909091 = 0.000164881 loss) | |
| I0506 01:18:34.879361 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:18:34.879381 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:18:34.879395 15760 solver.cpp:245] Train net output #149: total_confidence = 8.95553e-05 | |
| I0506 01:18:34.879406 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000371555 | |
| I0506 01:18:34.879420 15760 sgd_solver.cpp:106] Iteration 15000, lr = 0.001 | |
| I0506 01:18:50.323886 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.3662 > 30) by scale factor 0.802864 | |
| I0506 01:19:07.091051 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.7328 > 30) by scale factor 0.839566 | |
| I0506 01:19:32.420841 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.491 > 30) by scale factor 0.983898 | |
| I0506 01:20:22.258218 15760 solver.cpp:229] Iteration 15500, loss = 10.0185 | |
| I0506 01:20:22.258338 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.097561 | |
| I0506 01:20:22.258360 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:20:22.258373 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0506 01:20:22.258384 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.375 | |
| I0506 01:20:22.258396 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 01:20:22.258409 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:20:22.258420 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 01:20:22.258432 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 01:20:22.258445 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:20:22.258456 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:20:22.258467 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:20:22.258478 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:20:22.258491 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:20:22.258502 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:20:22.258512 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:20:22.258524 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:20:22.258536 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:20:22.258548 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:20:22.258559 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:20:22.258570 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:20:22.258582 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:20:22.258594 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:20:22.258605 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:20:22.258615 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.778409 | |
| I0506 01:20:22.258627 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.317073 | |
| I0506 01:20:22.258643 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.89251 (* 0.3 = 0.867752 loss) | |
| I0506 01:20:22.258657 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.831932 (* 0.3 = 0.24958 loss) | |
| I0506 01:20:22.258671 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.57678 (* 0.0272727 = 0.0702759 loss) | |
| I0506 01:20:22.258685 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.53107 (* 0.0272727 = 0.0963019 loss) | |
| I0506 01:20:22.258698 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 2.97432 (* 0.0272727 = 0.0811178 loss) | |
| I0506 01:20:22.258713 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.62921 (* 0.0272727 = 0.0989785 loss) | |
| I0506 01:20:22.258725 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.48679 (* 0.0272727 = 0.0678215 loss) | |
| I0506 01:20:22.258739 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.18862 (* 0.0272727 = 0.0596895 loss) | |
| I0506 01:20:22.258752 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.07173 (* 0.0272727 = 0.0292291 loss) | |
| I0506 01:20:22.258766 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.994714 (* 0.0272727 = 0.0271286 loss) | |
| I0506 01:20:22.258780 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.00521614 (* 0.0272727 = 0.000142258 loss) | |
| I0506 01:20:22.258795 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00396978 (* 0.0272727 = 0.000108267 loss) | |
| I0506 01:20:22.258808 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.00335319 (* 0.0272727 = 9.14505e-05 loss) | |
| I0506 01:20:22.258822 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00282798 (* 0.0272727 = 7.71267e-05 loss) | |
| I0506 01:20:22.258854 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00135551 (* 0.0272727 = 3.69684e-05 loss) | |
| I0506 01:20:22.258870 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00118944 (* 0.0272727 = 3.24392e-05 loss) | |
| I0506 01:20:22.258888 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00159376 (* 0.0272727 = 4.34663e-05 loss) | |
| I0506 01:20:22.258903 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00126895 (* 0.0272727 = 3.46077e-05 loss) | |
| I0506 01:20:22.258918 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.000901664 (* 0.0272727 = 2.45908e-05 loss) | |
| I0506 01:20:22.258931 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.000429478 (* 0.0272727 = 1.1713e-05 loss) | |
| I0506 01:20:22.258945 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.000731044 (* 0.0272727 = 1.99376e-05 loss) | |
| I0506 01:20:22.258958 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.000712979 (* 0.0272727 = 1.94449e-05 loss) | |
| I0506 01:20:22.258972 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.000401409 (* 0.0272727 = 1.09475e-05 loss) | |
| I0506 01:20:22.258985 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000517541 (* 0.0272727 = 1.41148e-05 loss) | |
| I0506 01:20:22.258997 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.146341 | |
| I0506 01:20:22.259009 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:20:22.259021 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:20:22.259033 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0506 01:20:22.259044 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 01:20:22.259055 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:20:22.259068 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 01:20:22.259078 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 01:20:22.259089 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 01:20:22.259101 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:20:22.259112 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:20:22.259124 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:20:22.259135 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:20:22.259145 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:20:22.259156 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:20:22.259167 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:20:22.259178 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:20:22.259189 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:20:22.259201 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:20:22.259212 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:20:22.259223 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:20:22.259234 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:20:22.259245 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:20:22.259256 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0506 01:20:22.259268 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.292683 | |
| I0506 01:20:22.259281 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.87106 (* 0.3 = 0.861319 loss) | |
| I0506 01:20:22.259294 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.921219 (* 0.3 = 0.276366 loss) | |
| I0506 01:20:22.259308 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.37139 (* 0.0272727 = 0.0646744 loss) | |
| I0506 01:20:22.259322 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.30616 (* 0.0272727 = 0.0901681 loss) | |
| I0506 01:20:22.259347 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 2.66005 (* 0.0272727 = 0.0725469 loss) | |
| I0506 01:20:22.259366 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.37073 (* 0.0272727 = 0.0919291 loss) | |
| I0506 01:20:22.259380 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.33933 (* 0.0272727 = 0.0638 loss) | |
| I0506 01:20:22.259393 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.51877 (* 0.0272727 = 0.041421 loss) | |
| I0506 01:20:22.259407 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.496957 (* 0.0272727 = 0.0135534 loss) | |
| I0506 01:20:22.259420 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.863992 (* 0.0272727 = 0.0235634 loss) | |
| I0506 01:20:22.259434 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00527777 (* 0.0272727 = 0.000143939 loss) | |
| I0506 01:20:22.259448 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00424488 (* 0.0272727 = 0.00011577 loss) | |
| I0506 01:20:22.259461 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00194418 (* 0.0272727 = 5.3023e-05 loss) | |
| I0506 01:20:22.259475 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00143735 (* 0.0272727 = 3.92003e-05 loss) | |
| I0506 01:20:22.259488 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00133623 (* 0.0272727 = 3.64428e-05 loss) | |
| I0506 01:20:22.259502 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.001927 (* 0.0272727 = 5.25544e-05 loss) | |
| I0506 01:20:22.259516 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00376584 (* 0.0272727 = 0.000102705 loss) | |
| I0506 01:20:22.259529 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.000297865 (* 0.0272727 = 8.1236e-06 loss) | |
| I0506 01:20:22.259543 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.000251314 (* 0.0272727 = 6.85403e-06 loss) | |
| I0506 01:20:22.259557 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.000592849 (* 0.0272727 = 1.61686e-05 loss) | |
| I0506 01:20:22.259570 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.000873099 (* 0.0272727 = 2.38118e-05 loss) | |
| I0506 01:20:22.259584 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.000245385 (* 0.0272727 = 6.69232e-06 loss) | |
| I0506 01:20:22.259598 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.000416082 (* 0.0272727 = 1.13477e-05 loss) | |
| I0506 01:20:22.259613 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.000439665 (* 0.0272727 = 1.19909e-05 loss) | |
| I0506 01:20:22.259624 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.121951 | |
| I0506 01:20:22.259635 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 01:20:22.259647 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:20:22.259659 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:20:22.259670 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 01:20:22.259680 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 01:20:22.259692 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 01:20:22.259703 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0506 01:20:22.259716 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 01:20:22.259727 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:20:22.259737 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:20:22.259748 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:20:22.259759 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:20:22.259770 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:20:22.259783 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:20:22.259793 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:20:22.259814 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:20:22.259826 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:20:22.259838 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:20:22.259850 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:20:22.259860 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:20:22.259871 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:20:22.259881 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:20:22.259892 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.778409 | |
| I0506 01:20:22.259904 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.390244 | |
| I0506 01:20:22.259917 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.70269 (* 1 = 2.70269 loss) | |
| I0506 01:20:22.259934 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.775542 (* 1 = 0.775542 loss) | |
| I0506 01:20:22.259949 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.20519 (* 0.0909091 = 0.200472 loss) | |
| I0506 01:20:22.259963 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.07664 (* 0.0909091 = 0.279695 loss) | |
| I0506 01:20:22.259976 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.92698 (* 0.0909091 = 0.266089 loss) | |
| I0506 01:20:22.259989 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.94695 (* 0.0909091 = 0.267904 loss) | |
| I0506 01:20:22.260002 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.74954 (* 0.0909091 = 0.159049 loss) | |
| I0506 01:20:22.260016 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.42449 (* 0.0909091 = 0.129499 loss) | |
| I0506 01:20:22.260026 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.563892 (* 0.0909091 = 0.0512629 loss) | |
| I0506 01:20:22.260035 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.637516 (* 0.0909091 = 0.057956 loss) | |
| I0506 01:20:22.260049 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.000987547 (* 0.0909091 = 8.9777e-05 loss) | |
| I0506 01:20:22.260063 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00042451 (* 0.0909091 = 3.85918e-05 loss) | |
| I0506 01:20:22.260076 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.000255228 (* 0.0909091 = 2.32026e-05 loss) | |
| I0506 01:20:22.260089 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.000254594 (* 0.0909091 = 2.31449e-05 loss) | |
| I0506 01:20:22.260103 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.000174291 (* 0.0909091 = 1.58447e-05 loss) | |
| I0506 01:20:22.260116 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000225521 (* 0.0909091 = 2.05019e-05 loss) | |
| I0506 01:20:22.260130 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00017171 (* 0.0909091 = 1.561e-05 loss) | |
| I0506 01:20:22.260143 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000210939 (* 0.0909091 = 1.91763e-05 loss) | |
| I0506 01:20:22.260157 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000152743 (* 0.0909091 = 1.38857e-05 loss) | |
| I0506 01:20:22.260170 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000250773 (* 0.0909091 = 2.27975e-05 loss) | |
| I0506 01:20:22.260185 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 9.55157e-05 (* 0.0909091 = 8.68325e-06 loss) | |
| I0506 01:20:22.260197 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 7.73192e-05 (* 0.0909091 = 7.02902e-06 loss) | |
| I0506 01:20:22.260211 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 8.03297e-05 (* 0.0909091 = 7.3027e-06 loss) | |
| I0506 01:20:22.260224 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 7.62979e-05 (* 0.0909091 = 6.93617e-06 loss) | |
| I0506 01:20:22.260236 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:20:22.260247 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:20:22.260267 15760 solver.cpp:245] Train net output #149: total_confidence = 7.34426e-06 | |
| I0506 01:20:22.260282 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000165033 | |
| I0506 01:20:22.260293 15760 sgd_solver.cpp:106] Iteration 15500, lr = 0.001 | |
| I0506 01:20:28.526556 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.369 > 30) by scale factor 0.956358 | |
| I0506 01:22:09.616576 15760 solver.cpp:229] Iteration 16000, loss = 10.0155 | |
| I0506 01:22:09.616736 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0588235 | |
| I0506 01:22:09.616757 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:22:09.616771 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:22:09.616782 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:22:09.616794 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:22:09.616807 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 01:22:09.616818 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:22:09.616830 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:22:09.616842 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.5 | |
| I0506 01:22:09.616854 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.5 | |
| I0506 01:22:09.616865 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.5 | |
| I0506 01:22:09.616880 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.625 | |
| I0506 01:22:09.616894 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.625 | |
| I0506 01:22:09.616905 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.75 | |
| I0506 01:22:09.616917 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 01:22:09.616930 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0506 01:22:09.616941 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0506 01:22:09.616953 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:22:09.616966 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:22:09.616977 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:22:09.616988 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:22:09.616999 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:22:09.617012 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:22:09.617022 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.585227 | |
| I0506 01:22:09.617034 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.235294 | |
| I0506 01:22:09.617051 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.30075 (* 0.3 = 0.990225 loss) | |
| I0506 01:22:09.617066 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.58378 (* 0.3 = 0.475135 loss) | |
| I0506 01:22:09.617080 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.29737 (* 0.0272727 = 0.0899282 loss) | |
| I0506 01:22:09.617094 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.55503 (* 0.0272727 = 0.0969554 loss) | |
| I0506 01:22:09.617108 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.55646 (* 0.0272727 = 0.0969944 loss) | |
| I0506 01:22:09.617138 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.97025 (* 0.0272727 = 0.0810069 loss) | |
| I0506 01:22:09.617154 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.30811 (* 0.0272727 = 0.0902213 loss) | |
| I0506 01:22:09.617169 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.08169 (* 0.0272727 = 0.0567733 loss) | |
| I0506 01:22:09.617182 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.10424 (* 0.0272727 = 0.0573884 loss) | |
| I0506 01:22:09.617197 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.99202 (* 0.0272727 = 0.0543277 loss) | |
| I0506 01:22:09.617209 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.62224 (* 0.0272727 = 0.0442429 loss) | |
| I0506 01:22:09.617223 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 1.78602 (* 0.0272727 = 0.0487096 loss) | |
| I0506 01:22:09.617236 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 1.53565 (* 0.0272727 = 0.0418814 loss) | |
| I0506 01:22:09.617250 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 1.72693 (* 0.0272727 = 0.0470982 loss) | |
| I0506 01:22:09.617283 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 1.25572 (* 0.0272727 = 0.0342469 loss) | |
| I0506 01:22:09.617298 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.650294 (* 0.0272727 = 0.0177353 loss) | |
| I0506 01:22:09.617312 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.709703 (* 0.0272727 = 0.0193555 loss) | |
| I0506 01:22:09.617326 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.59788 (* 0.0272727 = 0.0163058 loss) | |
| I0506 01:22:09.617341 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0541845 (* 0.0272727 = 0.00147776 loss) | |
| I0506 01:22:09.617355 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0444428 (* 0.0272727 = 0.00121208 loss) | |
| I0506 01:22:09.617369 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.023739 (* 0.0272727 = 0.000647428 loss) | |
| I0506 01:22:09.617383 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0389436 (* 0.0272727 = 0.0010621 loss) | |
| I0506 01:22:09.617398 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0277357 (* 0.0272727 = 0.000756428 loss) | |
| I0506 01:22:09.617411 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0218852 (* 0.0272727 = 0.000596868 loss) | |
| I0506 01:22:09.617424 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0441176 | |
| I0506 01:22:09.617435 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:22:09.617447 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:22:09.617460 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:22:09.617470 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 01:22:09.617482 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 01:22:09.617493 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0506 01:22:09.617506 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:22:09.617516 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.5 | |
| I0506 01:22:09.617528 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.5 | |
| I0506 01:22:09.617539 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.5 | |
| I0506 01:22:09.617552 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.625 | |
| I0506 01:22:09.617563 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.625 | |
| I0506 01:22:09.617573 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.75 | |
| I0506 01:22:09.617585 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 01:22:09.617597 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0506 01:22:09.617609 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0506 01:22:09.617622 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:22:09.617635 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:22:09.617645 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:22:09.617657 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:22:09.617669 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:22:09.617681 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:22:09.617691 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.630682 | |
| I0506 01:22:09.617703 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.220588 | |
| I0506 01:22:09.617717 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.17769 (* 0.3 = 0.953307 loss) | |
| I0506 01:22:09.617730 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.34911 (* 0.3 = 0.404734 loss) | |
| I0506 01:22:09.617748 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.19893 (* 0.0272727 = 0.0872436 loss) | |
| I0506 01:22:09.617763 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.18344 (* 0.0272727 = 0.0868211 loss) | |
| I0506 01:22:09.617787 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.22784 (* 0.0272727 = 0.0880321 loss) | |
| I0506 01:22:09.617802 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.3002 (* 0.0272727 = 0.0900055 loss) | |
| I0506 01:22:09.617816 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.01413 (* 0.0272727 = 0.0822035 loss) | |
| I0506 01:22:09.617830 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.99928 (* 0.0272727 = 0.0545258 loss) | |
| I0506 01:22:09.617843 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.18563 (* 0.0272727 = 0.059608 loss) | |
| I0506 01:22:09.617857 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.80454 (* 0.0272727 = 0.0492148 loss) | |
| I0506 01:22:09.617871 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.53695 (* 0.0272727 = 0.0419169 loss) | |
| I0506 01:22:09.617884 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 2.03682 (* 0.0272727 = 0.0555495 loss) | |
| I0506 01:22:09.617897 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 1.83435 (* 0.0272727 = 0.0500278 loss) | |
| I0506 01:22:09.617911 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 1.81815 (* 0.0272727 = 0.049586 loss) | |
| I0506 01:22:09.617928 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 1.43049 (* 0.0272727 = 0.0390133 loss) | |
| I0506 01:22:09.617943 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.555064 (* 0.0272727 = 0.0151381 loss) | |
| I0506 01:22:09.617956 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.731932 (* 0.0272727 = 0.0199618 loss) | |
| I0506 01:22:09.617970 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.607337 (* 0.0272727 = 0.0165637 loss) | |
| I0506 01:22:09.617983 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0624634 (* 0.0272727 = 0.00170355 loss) | |
| I0506 01:22:09.617997 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0532133 (* 0.0272727 = 0.00145127 loss) | |
| I0506 01:22:09.618011 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0298245 (* 0.0272727 = 0.000813396 loss) | |
| I0506 01:22:09.618024 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0263644 (* 0.0272727 = 0.00071903 loss) | |
| I0506 01:22:09.618038 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0368105 (* 0.0272727 = 0.00100392 loss) | |
| I0506 01:22:09.618052 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0143828 (* 0.0272727 = 0.000392259 loss) | |
| I0506 01:22:09.618063 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.117647 | |
| I0506 01:22:09.618075 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:22:09.618088 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:22:09.618098 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:22:09.618109 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.5 | |
| I0506 01:22:09.618121 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:22:09.618134 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 01:22:09.618144 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:22:09.618155 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.5 | |
| I0506 01:22:09.618167 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0506 01:22:09.618180 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.5 | |
| I0506 01:22:09.618191 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.625 | |
| I0506 01:22:09.618201 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.625 | |
| I0506 01:22:09.618213 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.75 | |
| I0506 01:22:09.618226 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 01:22:09.618237 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0506 01:22:09.618247 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0506 01:22:09.618268 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:22:09.618281 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:22:09.618294 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:22:09.618305 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:22:09.618316 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:22:09.618327 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:22:09.618338 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.659091 | |
| I0506 01:22:09.618350 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.279412 | |
| I0506 01:22:09.618363 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.94411 (* 1 = 2.94411 loss) | |
| I0506 01:22:09.618377 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.22606 (* 1 = 1.22606 loss) | |
| I0506 01:22:09.618391 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.88686 (* 0.0909091 = 0.262442 loss) | |
| I0506 01:22:09.618404 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.12062 (* 0.0909091 = 0.283693 loss) | |
| I0506 01:22:09.618418 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.07351 (* 0.0909091 = 0.27941 loss) | |
| I0506 01:22:09.618432 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.64689 (* 0.0909091 = 0.240626 loss) | |
| I0506 01:22:09.618444 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.76248 (* 0.0909091 = 0.251134 loss) | |
| I0506 01:22:09.618458 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.88796 (* 0.0909091 = 0.171633 loss) | |
| I0506 01:22:09.618471 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.0011 (* 0.0909091 = 0.181918 loss) | |
| I0506 01:22:09.618485 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.72197 (* 0.0909091 = 0.156543 loss) | |
| I0506 01:22:09.618499 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.56017 (* 0.0909091 = 0.141834 loss) | |
| I0506 01:22:09.618512 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 1.74588 (* 0.0909091 = 0.158717 loss) | |
| I0506 01:22:09.618525 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 1.39196 (* 0.0909091 = 0.126542 loss) | |
| I0506 01:22:09.618538 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 1.32776 (* 0.0909091 = 0.120706 loss) | |
| I0506 01:22:09.618552 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 1.11253 (* 0.0909091 = 0.101139 loss) | |
| I0506 01:22:09.618566 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.408071 (* 0.0909091 = 0.0370974 loss) | |
| I0506 01:22:09.618578 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.45507 (* 0.0909091 = 0.04137 loss) | |
| I0506 01:22:09.618592 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.47877 (* 0.0909091 = 0.0435246 loss) | |
| I0506 01:22:09.618605 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0191025 (* 0.0909091 = 0.00173659 loss) | |
| I0506 01:22:09.618619 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0154845 (* 0.0909091 = 0.00140768 loss) | |
| I0506 01:22:09.618633 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0139276 (* 0.0909091 = 0.00126615 loss) | |
| I0506 01:22:09.618646 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0144877 (* 0.0909091 = 0.00131706 loss) | |
| I0506 01:22:09.618660 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00595587 (* 0.0909091 = 0.000541443 loss) | |
| I0506 01:22:09.618674 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0041569 (* 0.0909091 = 0.0003779 loss) | |
| I0506 01:22:09.618686 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:22:09.618698 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:22:09.618710 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000200075 | |
| I0506 01:22:09.618729 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000509643 | |
| I0506 01:22:09.618744 15760 sgd_solver.cpp:106] Iteration 16000, lr = 0.001 | |
| I0506 01:22:52.101835 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 45.928 > 30) by scale factor 0.653196 | |
| I0506 01:23:56.735894 15760 solver.cpp:229] Iteration 16500, loss = 9.94345 | |
| I0506 01:23:56.736032 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.108696 | |
| I0506 01:23:56.736053 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:23:56.736066 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:23:56.736078 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:23:56.736090 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:23:56.736104 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:23:56.736114 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:23:56.736126 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:23:56.736138 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:23:56.736151 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:23:56.736162 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:23:56.736174 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:23:56.736186 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:23:56.736198 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:23:56.736210 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:23:56.736222 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:23:56.736233 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:23:56.736245 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:23:56.736258 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:23:56.736268 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:23:56.736280 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:23:56.736292 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:23:56.736304 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:23:56.736315 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0506 01:23:56.736326 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.217391 | |
| I0506 01:23:56.736343 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.95162 (* 0.3 = 0.885485 loss) | |
| I0506 01:23:56.736357 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.00708 (* 0.3 = 0.302125 loss) | |
| I0506 01:23:56.736371 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.68253 (* 0.0272727 = 0.0731598 loss) | |
| I0506 01:23:56.736385 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 2.70056 (* 0.0272727 = 0.0736517 loss) | |
| I0506 01:23:56.736398 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.42388 (* 0.0272727 = 0.0933785 loss) | |
| I0506 01:23:56.736413 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.80842 (* 0.0272727 = 0.0765933 loss) | |
| I0506 01:23:56.736425 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.3907 (* 0.0272727 = 0.065201 loss) | |
| I0506 01:23:56.736439 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.65103 (* 0.0272727 = 0.0723008 loss) | |
| I0506 01:23:56.736452 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.96038 (* 0.0272727 = 0.053465 loss) | |
| I0506 01:23:56.736466 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.863598 (* 0.0272727 = 0.0235527 loss) | |
| I0506 01:23:56.736480 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.381467 (* 0.0272727 = 0.0104036 loss) | |
| I0506 01:23:56.736493 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.594235 (* 0.0272727 = 0.0162064 loss) | |
| I0506 01:23:56.736508 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0420926 (* 0.0272727 = 0.00114798 loss) | |
| I0506 01:23:56.736522 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0222665 (* 0.0272727 = 0.000607269 loss) | |
| I0506 01:23:56.736555 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0186127 (* 0.0272727 = 0.00050762 loss) | |
| I0506 01:23:56.736570 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0114274 (* 0.0272727 = 0.000311657 loss) | |
| I0506 01:23:56.736585 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0104528 (* 0.0272727 = 0.000285076 loss) | |
| I0506 01:23:56.736599 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00780933 (* 0.0272727 = 0.000212982 loss) | |
| I0506 01:23:56.736613 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00562703 (* 0.0272727 = 0.000153464 loss) | |
| I0506 01:23:56.736626 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00611188 (* 0.0272727 = 0.000166688 loss) | |
| I0506 01:23:56.736640 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00314545 (* 0.0272727 = 8.57851e-05 loss) | |
| I0506 01:23:56.736654 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00338058 (* 0.0272727 = 9.21977e-05 loss) | |
| I0506 01:23:56.736668 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00204477 (* 0.0272727 = 5.57664e-05 loss) | |
| I0506 01:23:56.736682 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00446146 (* 0.0272727 = 0.000121676 loss) | |
| I0506 01:23:56.736695 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.130435 | |
| I0506 01:23:56.736706 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.375 | |
| I0506 01:23:56.736714 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:23:56.736722 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:23:56.736735 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 01:23:56.736747 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:23:56.736758 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:23:56.736770 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:23:56.736781 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 01:23:56.736793 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:23:56.736804 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:23:56.736816 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:23:56.736827 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:23:56.736838 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:23:56.736850 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:23:56.736860 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:23:56.736871 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:23:56.736886 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:23:56.736898 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:23:56.736910 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:23:56.736922 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:23:56.736932 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:23:56.736944 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:23:56.736955 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0506 01:23:56.736968 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.23913 | |
| I0506 01:23:56.736980 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.98506 (* 0.3 = 0.895518 loss) | |
| I0506 01:23:56.736994 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.91561 (* 0.3 = 0.274683 loss) | |
| I0506 01:23:56.737009 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.44143 (* 0.0272727 = 0.0665846 loss) | |
| I0506 01:23:56.737021 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.17463 (* 0.0272727 = 0.0865808 loss) | |
| I0506 01:23:56.737051 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.24347 (* 0.0272727 = 0.0884583 loss) | |
| I0506 01:23:56.737066 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.07812 (* 0.0272727 = 0.0839488 loss) | |
| I0506 01:23:56.737081 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.20907 (* 0.0272727 = 0.0602474 loss) | |
| I0506 01:23:56.737094 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.24959 (* 0.0272727 = 0.0613525 loss) | |
| I0506 01:23:56.737108 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.82522 (* 0.0272727 = 0.0497788 loss) | |
| I0506 01:23:56.737136 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.656586 (* 0.0272727 = 0.0179069 loss) | |
| I0506 01:23:56.737154 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.522952 (* 0.0272727 = 0.0142623 loss) | |
| I0506 01:23:56.737167 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.520556 (* 0.0272727 = 0.014197 loss) | |
| I0506 01:23:56.737181 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0377954 (* 0.0272727 = 0.00103078 loss) | |
| I0506 01:23:56.737195 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0202517 (* 0.0272727 = 0.000552319 loss) | |
| I0506 01:23:56.737208 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0151163 (* 0.0272727 = 0.000412262 loss) | |
| I0506 01:23:56.737222 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0101104 (* 0.0272727 = 0.000275737 loss) | |
| I0506 01:23:56.737236 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00799029 (* 0.0272727 = 0.000217917 loss) | |
| I0506 01:23:56.737249 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00551868 (* 0.0272727 = 0.000150509 loss) | |
| I0506 01:23:56.737263 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0017718 (* 0.0272727 = 4.83219e-05 loss) | |
| I0506 01:23:56.737278 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00290701 (* 0.0272727 = 7.92821e-05 loss) | |
| I0506 01:23:56.737293 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00313903 (* 0.0272727 = 8.56098e-05 loss) | |
| I0506 01:23:56.737305 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00285319 (* 0.0272727 = 7.78142e-05 loss) | |
| I0506 01:23:56.737319 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00745389 (* 0.0272727 = 0.000203288 loss) | |
| I0506 01:23:56.737334 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00342281 (* 0.0272727 = 9.33494e-05 loss) | |
| I0506 01:23:56.737345 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.173913 | |
| I0506 01:23:56.737356 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 01:23:56.737368 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:23:56.737380 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0506 01:23:56.737390 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:23:56.737401 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 01:23:56.737413 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:23:56.737424 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0506 01:23:56.737437 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 01:23:56.737448 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:23:56.737459 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:23:56.737471 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:23:56.737483 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:23:56.737493 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:23:56.737504 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:23:56.737515 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:23:56.737526 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:23:56.737550 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:23:56.737562 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:23:56.737573 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:23:56.737586 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:23:56.737596 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:23:56.737607 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:23:56.737618 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.778409 | |
| I0506 01:23:56.737630 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.347826 | |
| I0506 01:23:56.737643 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.75113 (* 1 = 2.75113 loss) | |
| I0506 01:23:56.737658 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.794765 (* 1 = 0.794765 loss) | |
| I0506 01:23:56.737670 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.42228 (* 0.0909091 = 0.220207 loss) | |
| I0506 01:23:56.737684 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.67373 (* 0.0909091 = 0.243066 loss) | |
| I0506 01:23:56.737697 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.82387 (* 0.0909091 = 0.256716 loss) | |
| I0506 01:23:56.737710 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.75818 (* 0.0909091 = 0.250744 loss) | |
| I0506 01:23:56.737725 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.82405 (* 0.0909091 = 0.165822 loss) | |
| I0506 01:23:56.737737 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.82471 (* 0.0909091 = 0.165883 loss) | |
| I0506 01:23:56.737751 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.4616 (* 0.0909091 = 0.132872 loss) | |
| I0506 01:23:56.737766 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.765494 (* 0.0909091 = 0.0695904 loss) | |
| I0506 01:23:56.737778 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.397768 (* 0.0909091 = 0.0361607 loss) | |
| I0506 01:23:56.737792 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.481252 (* 0.0909091 = 0.0437501 loss) | |
| I0506 01:23:56.737805 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0145984 (* 0.0909091 = 0.00132713 loss) | |
| I0506 01:23:56.737819 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0152466 (* 0.0909091 = 0.00138605 loss) | |
| I0506 01:23:56.737833 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0120426 (* 0.0909091 = 0.00109478 loss) | |
| I0506 01:23:56.737846 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00606218 (* 0.0909091 = 0.000551107 loss) | |
| I0506 01:23:56.737860 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00498137 (* 0.0909091 = 0.000452851 loss) | |
| I0506 01:23:56.737874 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00547663 (* 0.0909091 = 0.000497876 loss) | |
| I0506 01:23:56.737887 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00312826 (* 0.0909091 = 0.000284387 loss) | |
| I0506 01:23:56.737900 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00147396 (* 0.0909091 = 0.000133997 loss) | |
| I0506 01:23:56.737915 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00139995 (* 0.0909091 = 0.000127269 loss) | |
| I0506 01:23:56.737931 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00112971 (* 0.0909091 = 0.000102701 loss) | |
| I0506 01:23:56.737946 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000831991 (* 0.0909091 = 7.56355e-05 loss) | |
| I0506 01:23:56.737960 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000570404 (* 0.0909091 = 5.18549e-05 loss) | |
| I0506 01:23:56.737972 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:23:56.737983 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:23:56.738004 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000534827 | |
| I0506 01:23:56.738018 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000870216 | |
| I0506 01:23:56.738030 15760 sgd_solver.cpp:106] Iteration 16500, lr = 0.001 | |
| I0506 01:25:43.991717 15760 solver.cpp:229] Iteration 17000, loss = 9.93592 | |
| I0506 01:25:43.991858 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0677966 | |
| I0506 01:25:43.991880 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:25:43.991894 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:25:43.991907 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:25:43.991919 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:25:43.991930 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 01:25:43.991943 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:25:43.991955 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:25:43.991967 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 01:25:43.991979 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:25:43.991991 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:25:43.992003 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:25:43.992015 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:25:43.992034 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 01:25:43.992046 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 01:25:43.992058 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0506 01:25:43.992071 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0506 01:25:43.992084 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:25:43.992094 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:25:43.992106 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:25:43.992117 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:25:43.992130 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:25:43.992141 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:25:43.992151 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0506 01:25:43.992163 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.254237 | |
| I0506 01:25:43.992179 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.26894 (* 0.3 = 0.980682 loss) | |
| I0506 01:25:43.992193 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.17943 (* 0.3 = 0.353828 loss) | |
| I0506 01:25:43.992208 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.21518 (* 0.0272727 = 0.0876867 loss) | |
| I0506 01:25:43.992221 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.36198 (* 0.0272727 = 0.0916903 loss) | |
| I0506 01:25:43.992235 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.73144 (* 0.0272727 = 0.101766 loss) | |
| I0506 01:25:43.992249 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.43644 (* 0.0272727 = 0.0937212 loss) | |
| I0506 01:25:43.992262 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.12808 (* 0.0272727 = 0.0853114 loss) | |
| I0506 01:25:43.992276 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.72989 (* 0.0272727 = 0.0744517 loss) | |
| I0506 01:25:43.992290 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.86329 (* 0.0272727 = 0.0508169 loss) | |
| I0506 01:25:43.992310 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.74905 (* 0.0272727 = 0.0477013 loss) | |
| I0506 01:25:43.992323 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.72083 (* 0.0272727 = 0.019659 loss) | |
| I0506 01:25:43.992337 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.763149 (* 0.0272727 = 0.0208132 loss) | |
| I0506 01:25:43.992350 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.792021 (* 0.0272727 = 0.0216006 loss) | |
| I0506 01:25:43.992364 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.676367 (* 0.0272727 = 0.0184464 loss) | |
| I0506 01:25:43.992396 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.65164 (* 0.0272727 = 0.017772 loss) | |
| I0506 01:25:43.992411 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.595811 (* 0.0272727 = 0.0162494 loss) | |
| I0506 01:25:43.992425 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.850738 (* 0.0272727 = 0.0232019 loss) | |
| I0506 01:25:43.992439 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.782456 (* 0.0272727 = 0.0213397 loss) | |
| I0506 01:25:43.992454 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0295341 (* 0.0272727 = 0.000805475 loss) | |
| I0506 01:25:43.992466 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0175755 (* 0.0272727 = 0.000479331 loss) | |
| I0506 01:25:43.992480 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0120356 (* 0.0272727 = 0.000328245 loss) | |
| I0506 01:25:43.992494 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0195462 (* 0.0272727 = 0.000533079 loss) | |
| I0506 01:25:43.992508 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0130612 (* 0.0272727 = 0.000356214 loss) | |
| I0506 01:25:43.992522 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0115032 (* 0.0272727 = 0.000313724 loss) | |
| I0506 01:25:43.992533 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0338983 | |
| I0506 01:25:43.992545 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:25:43.992558 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:25:43.992568 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:25:43.992580 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 01:25:43.992591 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 01:25:43.992604 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:25:43.992615 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:25:43.992627 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0506 01:25:43.992638 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 01:25:43.992650 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:25:43.992662 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:25:43.992673 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:25:43.992686 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 01:25:43.992697 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 01:25:43.992708 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0506 01:25:43.992720 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0506 01:25:43.992733 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:25:43.992743 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:25:43.992754 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:25:43.992765 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:25:43.992776 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:25:43.992789 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:25:43.992799 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.659091 | |
| I0506 01:25:43.992811 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.254237 | |
| I0506 01:25:43.992825 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.3268 (* 0.3 = 0.998039 loss) | |
| I0506 01:25:43.992838 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.27955 (* 0.3 = 0.383866 loss) | |
| I0506 01:25:43.992851 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.06026 (* 0.0272727 = 0.0834616 loss) | |
| I0506 01:25:43.992866 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.49767 (* 0.0272727 = 0.095391 loss) | |
| I0506 01:25:43.992893 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.32879 (* 0.0272727 = 0.0907852 loss) | |
| I0506 01:25:43.992909 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.58122 (* 0.0272727 = 0.0976696 loss) | |
| I0506 01:25:43.992923 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.72277 (* 0.0272727 = 0.0742575 loss) | |
| I0506 01:25:43.992940 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.81279 (* 0.0272727 = 0.0767124 loss) | |
| I0506 01:25:43.992954 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.7841 (* 0.0272727 = 0.0486572 loss) | |
| I0506 01:25:43.992967 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 2.2305 (* 0.0272727 = 0.0608318 loss) | |
| I0506 01:25:43.992980 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.861565 (* 0.0272727 = 0.0234972 loss) | |
| I0506 01:25:43.992995 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.762614 (* 0.0272727 = 0.0207986 loss) | |
| I0506 01:25:43.993007 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.695635 (* 0.0272727 = 0.0189719 loss) | |
| I0506 01:25:43.993021 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.911969 (* 0.0272727 = 0.0248719 loss) | |
| I0506 01:25:43.993034 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.920529 (* 0.0272727 = 0.0251053 loss) | |
| I0506 01:25:43.993048 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.693535 (* 0.0272727 = 0.0189146 loss) | |
| I0506 01:25:43.993062 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.964285 (* 0.0272727 = 0.0262987 loss) | |
| I0506 01:25:43.993074 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.958382 (* 0.0272727 = 0.0261377 loss) | |
| I0506 01:25:43.993089 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.031243 (* 0.0272727 = 0.000852082 loss) | |
| I0506 01:25:43.993103 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0175685 (* 0.0272727 = 0.00047914 loss) | |
| I0506 01:25:43.993113 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0204817 (* 0.0272727 = 0.000558592 loss) | |
| I0506 01:25:43.993141 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0204225 (* 0.0272727 = 0.000556977 loss) | |
| I0506 01:25:43.993157 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.013855 (* 0.0272727 = 0.000377864 loss) | |
| I0506 01:25:43.993170 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0174338 (* 0.0272727 = 0.000475467 loss) | |
| I0506 01:25:43.993190 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.118644 | |
| I0506 01:25:43.993202 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:25:43.993214 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:25:43.993226 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:25:43.993237 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:25:43.993248 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:25:43.993260 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:25:43.993271 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:25:43.993283 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0506 01:25:43.993294 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 01:25:43.993305 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:25:43.993316 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:25:43.993327 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:25:43.993340 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 01:25:43.993350 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 01:25:43.993361 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0506 01:25:43.993372 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0506 01:25:43.993394 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:25:43.993407 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:25:43.993418 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:25:43.993429 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:25:43.993440 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:25:43.993453 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:25:43.993463 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.681818 | |
| I0506 01:25:43.993474 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.38983 | |
| I0506 01:25:43.993487 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.97575 (* 1 = 2.97575 loss) | |
| I0506 01:25:43.993500 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.15731 (* 1 = 1.15731 loss) | |
| I0506 01:25:43.993515 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.89693 (* 0.0909091 = 0.263357 loss) | |
| I0506 01:25:43.993527 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.09933 (* 0.0909091 = 0.281757 loss) | |
| I0506 01:25:43.993541 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.56477 (* 0.0909091 = 0.32407 loss) | |
| I0506 01:25:43.993553 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.20304 (* 0.0909091 = 0.291185 loss) | |
| I0506 01:25:43.993566 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.58208 (* 0.0909091 = 0.234735 loss) | |
| I0506 01:25:43.993579 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.35369 (* 0.0909091 = 0.213971 loss) | |
| I0506 01:25:43.993593 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.72875 (* 0.0909091 = 0.157159 loss) | |
| I0506 01:25:43.993607 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.56987 (* 0.0909091 = 0.142715 loss) | |
| I0506 01:25:43.993619 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.516857 (* 0.0909091 = 0.046987 loss) | |
| I0506 01:25:43.993633 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.565294 (* 0.0909091 = 0.0513904 loss) | |
| I0506 01:25:43.993646 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.52204 (* 0.0909091 = 0.0474582 loss) | |
| I0506 01:25:43.993659 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.586505 (* 0.0909091 = 0.0533186 loss) | |
| I0506 01:25:43.993674 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.458763 (* 0.0909091 = 0.0417057 loss) | |
| I0506 01:25:43.993686 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.515217 (* 0.0909091 = 0.0468379 loss) | |
| I0506 01:25:43.993700 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.502176 (* 0.0909091 = 0.0456524 loss) | |
| I0506 01:25:43.993712 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.624175 (* 0.0909091 = 0.0567432 loss) | |
| I0506 01:25:43.993726 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0138602 (* 0.0909091 = 0.00126001 loss) | |
| I0506 01:25:43.993741 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0191622 (* 0.0909091 = 0.00174201 loss) | |
| I0506 01:25:43.993753 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0130351 (* 0.0909091 = 0.00118501 loss) | |
| I0506 01:25:43.993767 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00799869 (* 0.0909091 = 0.000727154 loss) | |
| I0506 01:25:43.993780 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00593357 (* 0.0909091 = 0.000539415 loss) | |
| I0506 01:25:43.993794 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00304111 (* 0.0909091 = 0.000276465 loss) | |
| I0506 01:25:43.993806 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:25:43.993818 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:25:43.993829 15760 solver.cpp:245] Train net output #149: total_confidence = 2.62953e-05 | |
| I0506 01:25:43.993849 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000133172 | |
| I0506 01:25:43.993863 15760 sgd_solver.cpp:106] Iteration 17000, lr = 0.001 | |
| I0506 01:27:31.295233 15760 solver.cpp:229] Iteration 17500, loss = 9.92399 | |
| I0506 01:27:31.295379 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.148936 | |
| I0506 01:27:31.295400 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:27:31.295413 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:27:31.295425 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:27:31.295438 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:27:31.295449 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:27:31.295461 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:27:31.295474 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:27:31.295485 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0506 01:27:31.295497 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:27:31.295509 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:27:31.295521 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:27:31.295532 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:27:31.295544 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:27:31.295555 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:27:31.295567 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:27:31.295579 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:27:31.295591 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:27:31.295603 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:27:31.295615 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:27:31.295626 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:27:31.295639 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:27:31.295650 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:27:31.295661 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0506 01:27:31.295673 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.234043 | |
| I0506 01:27:31.295691 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.28264 (* 0.3 = 0.984792 loss) | |
| I0506 01:27:31.295704 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.31676 (* 0.3 = 0.395028 loss) | |
| I0506 01:27:31.295727 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.37559 (* 0.0272727 = 0.0920615 loss) | |
| I0506 01:27:31.295742 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.15503 (* 0.0272727 = 0.0860463 loss) | |
| I0506 01:27:31.295755 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.58466 (* 0.0272727 = 0.0977635 loss) | |
| I0506 01:27:31.295769 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.55164 (* 0.0272727 = 0.0968628 loss) | |
| I0506 01:27:31.295783 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.95932 (* 0.0272727 = 0.0807087 loss) | |
| I0506 01:27:31.295797 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.62397 (* 0.0272727 = 0.0715628 loss) | |
| I0506 01:27:31.295810 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.96144 (* 0.0272727 = 0.0534938 loss) | |
| I0506 01:27:31.295825 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.24654 (* 0.0272727 = 0.0339965 loss) | |
| I0506 01:27:31.295837 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.818099 (* 0.0272727 = 0.0223118 loss) | |
| I0506 01:27:31.295852 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.333367 (* 0.0272727 = 0.00909184 loss) | |
| I0506 01:27:31.295866 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.282768 (* 0.0272727 = 0.00771185 loss) | |
| I0506 01:27:31.295883 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.218396 (* 0.0272727 = 0.00595626 loss) | |
| I0506 01:27:31.295898 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.125067 (* 0.0272727 = 0.00341092 loss) | |
| I0506 01:27:31.295933 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.103732 (* 0.0272727 = 0.00282906 loss) | |
| I0506 01:27:31.295948 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0887034 (* 0.0272727 = 0.00241918 loss) | |
| I0506 01:27:31.295961 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0417917 (* 0.0272727 = 0.00113977 loss) | |
| I0506 01:27:31.295975 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0316902 (* 0.0272727 = 0.000864277 loss) | |
| I0506 01:27:31.295989 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0206065 (* 0.0272727 = 0.000561995 loss) | |
| I0506 01:27:31.296003 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0132933 (* 0.0272727 = 0.000362544 loss) | |
| I0506 01:27:31.296017 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0172034 (* 0.0272727 = 0.000469184 loss) | |
| I0506 01:27:31.296030 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.012377 (* 0.0272727 = 0.000337553 loss) | |
| I0506 01:27:31.296044 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0212143 (* 0.0272727 = 0.000578572 loss) | |
| I0506 01:27:31.296056 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.148936 | |
| I0506 01:27:31.296068 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:27:31.296080 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:27:31.296092 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:27:31.296104 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 01:27:31.296115 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 01:27:31.296128 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:27:31.296139 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:27:31.296150 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:27:31.296161 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 01:27:31.296174 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:27:31.296185 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:27:31.296196 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:27:31.296207 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:27:31.296218 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:27:31.296231 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:27:31.296241 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:27:31.296252 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:27:31.296263 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:27:31.296275 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:27:31.296286 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:27:31.296298 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:27:31.296309 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:27:31.296317 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.704545 | |
| I0506 01:27:31.296325 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.276596 | |
| I0506 01:27:31.296335 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.19658 (* 0.3 = 0.958975 loss) | |
| I0506 01:27:31.296352 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.35875 (* 0.3 = 0.407626 loss) | |
| I0506 01:27:31.296366 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.0947 (* 0.0272727 = 0.0844009 loss) | |
| I0506 01:27:31.296380 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.46317 (* 0.0272727 = 0.0944502 loss) | |
| I0506 01:27:31.296416 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.15938 (* 0.0272727 = 0.0861648 loss) | |
| I0506 01:27:31.296430 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.7172 (* 0.0272727 = 0.101378 loss) | |
| I0506 01:27:31.296444 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.12332 (* 0.0272727 = 0.0851815 loss) | |
| I0506 01:27:31.296458 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.7075 (* 0.0272727 = 0.0738408 loss) | |
| I0506 01:27:31.296473 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.8352 (* 0.0272727 = 0.050051 loss) | |
| I0506 01:27:31.296485 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.13909 (* 0.0272727 = 0.0310662 loss) | |
| I0506 01:27:31.296499 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.04301 (* 0.0272727 = 0.0284457 loss) | |
| I0506 01:27:31.296514 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.302543 (* 0.0272727 = 0.00825116 loss) | |
| I0506 01:27:31.296526 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.256308 (* 0.0272727 = 0.00699022 loss) | |
| I0506 01:27:31.296540 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.175673 (* 0.0272727 = 0.00479108 loss) | |
| I0506 01:27:31.296555 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.11875 (* 0.0272727 = 0.00323864 loss) | |
| I0506 01:27:31.296567 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0579148 (* 0.0272727 = 0.00157949 loss) | |
| I0506 01:27:31.296581 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0405887 (* 0.0272727 = 0.00110696 loss) | |
| I0506 01:27:31.296596 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0204879 (* 0.0272727 = 0.000558762 loss) | |
| I0506 01:27:31.296608 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00865761 (* 0.0272727 = 0.000236117 loss) | |
| I0506 01:27:31.296622 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00808234 (* 0.0272727 = 0.000220427 loss) | |
| I0506 01:27:31.296635 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00488437 (* 0.0272727 = 0.00013321 loss) | |
| I0506 01:27:31.296649 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00532533 (* 0.0272727 = 0.000145236 loss) | |
| I0506 01:27:31.296663 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00468999 (* 0.0272727 = 0.000127909 loss) | |
| I0506 01:27:31.296676 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0041603 (* 0.0272727 = 0.000113463 loss) | |
| I0506 01:27:31.296689 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.12766 | |
| I0506 01:27:31.296700 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:27:31.296711 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:27:31.296722 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.375 | |
| I0506 01:27:31.296735 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:27:31.296746 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:27:31.296756 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 01:27:31.296768 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:27:31.296779 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:27:31.296790 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:27:31.296802 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:27:31.296813 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:27:31.296824 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:27:31.296835 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:27:31.296847 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:27:31.296859 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:27:31.296869 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:27:31.296890 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:27:31.296903 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:27:31.296916 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:27:31.296929 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:27:31.296941 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:27:31.296953 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:27:31.296964 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0506 01:27:31.296977 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.276596 | |
| I0506 01:27:31.296990 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.07685 (* 1 = 3.07685 loss) | |
| I0506 01:27:31.297003 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.03974 (* 1 = 1.03974 loss) | |
| I0506 01:27:31.297018 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.80935 (* 0.0909091 = 0.255395 loss) | |
| I0506 01:27:31.297030 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.02275 (* 0.0909091 = 0.274795 loss) | |
| I0506 01:27:31.297044 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.80305 (* 0.0909091 = 0.254823 loss) | |
| I0506 01:27:31.297057 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.21661 (* 0.0909091 = 0.292419 loss) | |
| I0506 01:27:31.297071 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.66864 (* 0.0909091 = 0.242603 loss) | |
| I0506 01:27:31.297085 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.21453 (* 0.0909091 = 0.201321 loss) | |
| I0506 01:27:31.297097 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.63165 (* 0.0909091 = 0.148332 loss) | |
| I0506 01:27:31.297111 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.979778 (* 0.0909091 = 0.0890708 loss) | |
| I0506 01:27:31.297143 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.17348 (* 0.0909091 = 0.10668 loss) | |
| I0506 01:27:31.297159 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.16045 (* 0.0909091 = 0.0145864 loss) | |
| I0506 01:27:31.297173 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.183888 (* 0.0909091 = 0.0167171 loss) | |
| I0506 01:27:31.297186 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0860582 (* 0.0909091 = 0.00782348 loss) | |
| I0506 01:27:31.297200 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0438154 (* 0.0909091 = 0.00398322 loss) | |
| I0506 01:27:31.297214 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0478049 (* 0.0909091 = 0.0043459 loss) | |
| I0506 01:27:31.297227 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0165837 (* 0.0909091 = 0.00150761 loss) | |
| I0506 01:27:31.297241 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00926764 (* 0.0909091 = 0.000842513 loss) | |
| I0506 01:27:31.297255 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00228216 (* 0.0909091 = 0.000207469 loss) | |
| I0506 01:27:31.297269 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00182974 (* 0.0909091 = 0.00016634 loss) | |
| I0506 01:27:31.297283 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00107091 (* 0.0909091 = 9.73553e-05 loss) | |
| I0506 01:27:31.297297 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000705886 (* 0.0909091 = 6.41715e-05 loss) | |
| I0506 01:27:31.297312 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000643947 (* 0.0909091 = 5.85407e-05 loss) | |
| I0506 01:27:31.297324 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00041101 (* 0.0909091 = 3.73646e-05 loss) | |
| I0506 01:27:31.297336 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:27:31.297348 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:27:31.297358 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000256744 | |
| I0506 01:27:31.297381 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000820921 | |
| I0506 01:27:31.297395 15760 sgd_solver.cpp:106] Iteration 17500, lr = 0.001 | |
| I0506 01:29:01.606556 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.5051 > 30) by scale factor 0.869437 | |
| I0506 01:29:18.691413 15760 solver.cpp:229] Iteration 18000, loss = 9.84151 | |
| I0506 01:29:18.691473 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.102564 | |
| I0506 01:29:18.691489 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:29:18.691503 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:29:18.691514 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:29:18.691526 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:29:18.691539 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.75 | |
| I0506 01:29:18.691550 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 01:29:18.691563 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 01:29:18.691576 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 01:29:18.691587 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:29:18.691598 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:29:18.691611 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:29:18.691622 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:29:18.691633 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:29:18.691645 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:29:18.691658 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:29:18.691669 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:29:18.691681 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:29:18.691694 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:29:18.691705 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:29:18.691716 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:29:18.691728 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:29:18.691740 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:29:18.691751 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.795455 | |
| I0506 01:29:18.691763 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.333333 | |
| I0506 01:29:18.691779 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.904 (* 0.3 = 0.871201 loss) | |
| I0506 01:29:18.691794 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.745958 (* 0.3 = 0.223787 loss) | |
| I0506 01:29:18.691808 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.72143 (* 0.0272727 = 0.0742207 loss) | |
| I0506 01:29:18.691823 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.06869 (* 0.0272727 = 0.0836914 loss) | |
| I0506 01:29:18.691839 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.29165 (* 0.0272727 = 0.0897723 loss) | |
| I0506 01:29:18.691854 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.32526 (* 0.0272727 = 0.0906888 loss) | |
| I0506 01:29:18.691867 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 1.66562 (* 0.0272727 = 0.0454259 loss) | |
| I0506 01:29:18.691881 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.49849 (* 0.0272727 = 0.040868 loss) | |
| I0506 01:29:18.691895 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.2468 (* 0.0272727 = 0.0340037 loss) | |
| I0506 01:29:18.691910 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.157636 (* 0.0272727 = 0.00429915 loss) | |
| I0506 01:29:18.691923 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0265715 (* 0.0272727 = 0.000724678 loss) | |
| I0506 01:29:18.691937 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0322698 (* 0.0272727 = 0.000880085 loss) | |
| I0506 01:29:18.691951 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0143936 (* 0.0272727 = 0.000392553 loss) | |
| I0506 01:29:18.691967 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0142723 (* 0.0272727 = 0.000389244 loss) | |
| I0506 01:29:18.692013 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0112115 (* 0.0272727 = 0.000305769 loss) | |
| I0506 01:29:18.692029 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00858928 (* 0.0272727 = 0.000234253 loss) | |
| I0506 01:29:18.692044 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00661772 (* 0.0272727 = 0.000180483 loss) | |
| I0506 01:29:18.692057 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00611934 (* 0.0272727 = 0.000166891 loss) | |
| I0506 01:29:18.692071 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00367219 (* 0.0272727 = 0.000100151 loss) | |
| I0506 01:29:18.692085 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00307575 (* 0.0272727 = 8.38842e-05 loss) | |
| I0506 01:29:18.692100 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00503934 (* 0.0272727 = 0.000137436 loss) | |
| I0506 01:29:18.692113 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00204089 (* 0.0272727 = 5.56606e-05 loss) | |
| I0506 01:29:18.692126 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00224598 (* 0.0272727 = 6.12541e-05 loss) | |
| I0506 01:29:18.692140 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00179553 (* 0.0272727 = 4.89691e-05 loss) | |
| I0506 01:29:18.692153 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.102564 | |
| I0506 01:29:18.692165 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0506 01:29:18.692178 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:29:18.692189 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:29:18.692200 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:29:18.692212 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.625 | |
| I0506 01:29:18.692224 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 01:29:18.692236 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 01:29:18.692247 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 01:29:18.692260 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:29:18.692270 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:29:18.692286 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:29:18.692298 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:29:18.692309 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:29:18.692322 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:29:18.692332 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:29:18.692344 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:29:18.692355 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:29:18.692366 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:29:18.692378 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:29:18.692389 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:29:18.692401 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:29:18.692412 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:29:18.692425 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.778409 | |
| I0506 01:29:18.692435 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.230769 | |
| I0506 01:29:18.692450 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.88019 (* 0.3 = 0.864056 loss) | |
| I0506 01:29:18.692463 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.757471 (* 0.3 = 0.227241 loss) | |
| I0506 01:29:18.692477 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.61931 (* 0.0272727 = 0.0714358 loss) | |
| I0506 01:29:18.692492 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.13169 (* 0.0272727 = 0.0854099 loss) | |
| I0506 01:29:18.692517 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 2.8764 (* 0.0272727 = 0.0784472 loss) | |
| I0506 01:29:18.692530 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.15442 (* 0.0272727 = 0.0860296 loss) | |
| I0506 01:29:18.692544 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 1.52634 (* 0.0272727 = 0.0416275 loss) | |
| I0506 01:29:18.692559 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.36576 (* 0.0272727 = 0.0372481 loss) | |
| I0506 01:29:18.692571 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.14483 (* 0.0272727 = 0.0312225 loss) | |
| I0506 01:29:18.692585 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0348468 (* 0.0272727 = 0.000950367 loss) | |
| I0506 01:29:18.692600 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00318891 (* 0.0272727 = 8.69704e-05 loss) | |
| I0506 01:29:18.692613 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0020861 (* 0.0272727 = 5.68935e-05 loss) | |
| I0506 01:29:18.692627 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.000919052 (* 0.0272727 = 2.50651e-05 loss) | |
| I0506 01:29:18.692641 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.000775888 (* 0.0272727 = 2.11606e-05 loss) | |
| I0506 01:29:18.692654 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00062808 (* 0.0272727 = 1.71295e-05 loss) | |
| I0506 01:29:18.692668 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.000446156 (* 0.0272727 = 1.21679e-05 loss) | |
| I0506 01:29:18.692682 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.000765015 (* 0.0272727 = 2.08641e-05 loss) | |
| I0506 01:29:18.692695 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.000873215 (* 0.0272727 = 2.38149e-05 loss) | |
| I0506 01:29:18.692709 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.000243824 (* 0.0272727 = 6.64976e-06 loss) | |
| I0506 01:29:18.692723 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.000120272 (* 0.0272727 = 3.28014e-06 loss) | |
| I0506 01:29:18.692737 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.000203077 (* 0.0272727 = 5.53847e-06 loss) | |
| I0506 01:29:18.692751 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.000128699 (* 0.0272727 = 3.50997e-06 loss) | |
| I0506 01:29:18.692765 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.000362223 (* 0.0272727 = 9.8788e-06 loss) | |
| I0506 01:29:18.692778 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 5.3572e-05 (* 0.0272727 = 1.46105e-06 loss) | |
| I0506 01:29:18.692790 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.128205 | |
| I0506 01:29:18.692802 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:29:18.692814 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 01:29:18.692826 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0506 01:29:18.692837 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:29:18.692848 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.75 | |
| I0506 01:29:18.692859 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 01:29:18.692870 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 01:29:18.692883 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 01:29:18.692898 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:29:18.692909 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:29:18.692920 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:29:18.692931 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:29:18.692942 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:29:18.692953 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:29:18.692965 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:29:18.692986 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:29:18.692998 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:29:18.693009 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:29:18.693020 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:29:18.693032 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:29:18.693043 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:29:18.693054 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:29:18.693065 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.795455 | |
| I0506 01:29:18.693078 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.410256 | |
| I0506 01:29:18.693091 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.78117 (* 1 = 2.78117 loss) | |
| I0506 01:29:18.693101 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.693835 (* 1 = 0.693835 loss) | |
| I0506 01:29:18.693115 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.56413 (* 0.0909091 = 0.233103 loss) | |
| I0506 01:29:18.693145 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.61322 (* 0.0909091 = 0.237565 loss) | |
| I0506 01:29:18.693158 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.571 (* 0.0909091 = 0.233727 loss) | |
| I0506 01:29:18.693172 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.95808 (* 0.0909091 = 0.268917 loss) | |
| I0506 01:29:18.693186 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.07135 (* 0.0909091 = 0.0973955 loss) | |
| I0506 01:29:18.693198 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.37143 (* 0.0909091 = 0.124675 loss) | |
| I0506 01:29:18.693212 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.03956 (* 0.0909091 = 0.0945052 loss) | |
| I0506 01:29:18.693227 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0834733 (* 0.0909091 = 0.00758848 loss) | |
| I0506 01:29:18.693239 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00435222 (* 0.0909091 = 0.000395657 loss) | |
| I0506 01:29:18.693253 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00304929 (* 0.0909091 = 0.000277209 loss) | |
| I0506 01:29:18.693266 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00237025 (* 0.0909091 = 0.000215477 loss) | |
| I0506 01:29:18.693279 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00196659 (* 0.0909091 = 0.000178781 loss) | |
| I0506 01:29:18.693294 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00142652 (* 0.0909091 = 0.000129683 loss) | |
| I0506 01:29:18.693306 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0015665 (* 0.0909091 = 0.000142409 loss) | |
| I0506 01:29:18.693320 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000896199 (* 0.0909091 = 8.14727e-05 loss) | |
| I0506 01:29:18.693337 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00101171 (* 0.0909091 = 9.19733e-05 loss) | |
| I0506 01:29:18.693351 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000667271 (* 0.0909091 = 6.0661e-05 loss) | |
| I0506 01:29:18.693366 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000944585 (* 0.0909091 = 8.58713e-05 loss) | |
| I0506 01:29:18.693379 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000604601 (* 0.0909091 = 5.49638e-05 loss) | |
| I0506 01:29:18.693392 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000727181 (* 0.0909091 = 6.61074e-05 loss) | |
| I0506 01:29:18.693406 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000259394 (* 0.0909091 = 2.35813e-05 loss) | |
| I0506 01:29:18.693419 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000264789 (* 0.0909091 = 2.40717e-05 loss) | |
| I0506 01:29:18.693431 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:29:18.693454 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:29:18.693472 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000242945 | |
| I0506 01:29:18.693496 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000672208 | |
| I0506 01:29:18.693514 15760 sgd_solver.cpp:106] Iteration 18000, lr = 0.001 | |
| I0506 01:30:07.061231 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.7386 > 30) by scale factor 0.975973 | |
| I0506 01:31:05.965481 15760 solver.cpp:229] Iteration 18500, loss = 9.96795 | |
| I0506 01:31:05.965617 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0943396 | |
| I0506 01:31:05.965638 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:31:05.965651 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:31:05.965663 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:31:05.965675 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:31:05.965687 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 01:31:05.965699 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 01:31:05.965710 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:31:05.965723 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:31:05.965734 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:31:05.965746 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:31:05.965759 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:31:05.965770 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:31:05.965782 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:31:05.965793 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:31:05.965806 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:31:05.965816 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:31:05.965828 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:31:05.965840 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:31:05.965852 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:31:05.965862 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:31:05.965878 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:31:05.965889 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:31:05.965910 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.727273 | |
| I0506 01:31:05.965936 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.207547 | |
| I0506 01:31:05.965955 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.07283 (* 0.3 = 0.921848 loss) | |
| I0506 01:31:05.965968 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.968718 (* 0.3 = 0.290615 loss) | |
| I0506 01:31:05.965982 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.09016 (* 0.0272727 = 0.0842771 loss) | |
| I0506 01:31:05.965996 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.06972 (* 0.0272727 = 0.0837197 loss) | |
| I0506 01:31:05.966009 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.21976 (* 0.0272727 = 0.0878115 loss) | |
| I0506 01:31:05.966023 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.86161 (* 0.0272727 = 0.0780438 loss) | |
| I0506 01:31:05.966037 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.646 (* 0.0272727 = 0.0721637 loss) | |
| I0506 01:31:05.966050 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.73375 (* 0.0272727 = 0.0745567 loss) | |
| I0506 01:31:05.966064 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.15323 (* 0.0272727 = 0.0587244 loss) | |
| I0506 01:31:05.966078 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.770748 (* 0.0272727 = 0.0210204 loss) | |
| I0506 01:31:05.966091 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.667598 (* 0.0272727 = 0.0182072 loss) | |
| I0506 01:31:05.966105 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.567386 (* 0.0272727 = 0.0154742 loss) | |
| I0506 01:31:05.966119 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.638171 (* 0.0272727 = 0.0174047 loss) | |
| I0506 01:31:05.966132 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.518874 (* 0.0272727 = 0.0141511 loss) | |
| I0506 01:31:05.966146 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.20874 (* 0.0272727 = 0.00569292 loss) | |
| I0506 01:31:05.966181 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.145455 (* 0.0272727 = 0.00396696 loss) | |
| I0506 01:31:05.966195 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.132394 (* 0.0272727 = 0.00361075 loss) | |
| I0506 01:31:05.966210 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0941058 (* 0.0272727 = 0.00256652 loss) | |
| I0506 01:31:05.966224 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0479438 (* 0.0272727 = 0.00130756 loss) | |
| I0506 01:31:05.966238 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0213675 (* 0.0272727 = 0.000582749 loss) | |
| I0506 01:31:05.966253 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0170765 (* 0.0272727 = 0.000465723 loss) | |
| I0506 01:31:05.966265 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0143635 (* 0.0272727 = 0.000391733 loss) | |
| I0506 01:31:05.966279 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0118474 (* 0.0272727 = 0.00032311 loss) | |
| I0506 01:31:05.966294 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0138069 (* 0.0272727 = 0.000376552 loss) | |
| I0506 01:31:05.966305 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0943396 | |
| I0506 01:31:05.966317 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:31:05.966330 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:31:05.966341 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:31:05.966351 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:31:05.966363 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 01:31:05.966374 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0506 01:31:05.966387 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:31:05.966398 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 01:31:05.966409 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:31:05.966421 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:31:05.966429 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:31:05.966437 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:31:05.966450 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:31:05.966471 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:31:05.966485 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:31:05.966505 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:31:05.966517 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:31:05.966528 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:31:05.966541 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:31:05.966552 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:31:05.966562 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:31:05.966574 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:31:05.966585 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.721591 | |
| I0506 01:31:05.966598 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.226415 | |
| I0506 01:31:05.966611 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.11192 (* 0.3 = 0.933577 loss) | |
| I0506 01:31:05.966629 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.04694 (* 0.3 = 0.314083 loss) | |
| I0506 01:31:05.966645 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.0313 (* 0.0272727 = 0.0826719 loss) | |
| I0506 01:31:05.966658 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.95098 (* 0.0272727 = 0.0804814 loss) | |
| I0506 01:31:05.966684 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.30574 (* 0.0272727 = 0.0901565 loss) | |
| I0506 01:31:05.966699 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.07955 (* 0.0272727 = 0.0839877 loss) | |
| I0506 01:31:05.966713 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.07402 (* 0.0272727 = 0.0838369 loss) | |
| I0506 01:31:05.966727 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.87024 (* 0.0272727 = 0.0782793 loss) | |
| I0506 01:31:05.966740 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.3266 (* 0.0272727 = 0.0634528 loss) | |
| I0506 01:31:05.966753 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.696118 (* 0.0272727 = 0.018985 loss) | |
| I0506 01:31:05.966768 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.577623 (* 0.0272727 = 0.0157533 loss) | |
| I0506 01:31:05.966781 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.690363 (* 0.0272727 = 0.0188281 loss) | |
| I0506 01:31:05.966795 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.512631 (* 0.0272727 = 0.0139808 loss) | |
| I0506 01:31:05.966809 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.451186 (* 0.0272727 = 0.0123051 loss) | |
| I0506 01:31:05.966822 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.207339 (* 0.0272727 = 0.00565469 loss) | |
| I0506 01:31:05.966837 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.126093 (* 0.0272727 = 0.0034389 loss) | |
| I0506 01:31:05.966850 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0598614 (* 0.0272727 = 0.00163258 loss) | |
| I0506 01:31:05.966864 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0788703 (* 0.0272727 = 0.00215101 loss) | |
| I0506 01:31:05.966878 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0350391 (* 0.0272727 = 0.000955613 loss) | |
| I0506 01:31:05.966892 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0125406 (* 0.0272727 = 0.000342017 loss) | |
| I0506 01:31:05.966905 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0151844 (* 0.0272727 = 0.000414119 loss) | |
| I0506 01:31:05.966919 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0235293 (* 0.0272727 = 0.000641709 loss) | |
| I0506 01:31:05.966935 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0168701 (* 0.0272727 = 0.000460093 loss) | |
| I0506 01:31:05.966950 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0114119 (* 0.0272727 = 0.000311235 loss) | |
| I0506 01:31:05.966963 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0377358 | |
| I0506 01:31:05.966974 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:31:05.966985 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:31:05.966996 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:31:05.967008 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:31:05.967020 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:31:05.967031 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 01:31:05.967042 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:31:05.967053 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 01:31:05.967066 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:31:05.967077 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:31:05.967087 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:31:05.967099 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:31:05.967110 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:31:05.967123 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:31:05.967133 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:31:05.967144 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:31:05.967164 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:31:05.967177 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:31:05.967190 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:31:05.967200 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:31:05.967211 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:31:05.967222 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:31:05.967233 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0506 01:31:05.967245 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.207547 | |
| I0506 01:31:05.967258 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.92021 (* 1 = 2.92021 loss) | |
| I0506 01:31:05.967272 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.998333 (* 1 = 0.998333 loss) | |
| I0506 01:31:05.967286 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.7211 (* 0.0909091 = 0.247373 loss) | |
| I0506 01:31:05.967300 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.87189 (* 0.0909091 = 0.261081 loss) | |
| I0506 01:31:05.967314 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.77382 (* 0.0909091 = 0.252165 loss) | |
| I0506 01:31:05.967326 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.77372 (* 0.0909091 = 0.252156 loss) | |
| I0506 01:31:05.967340 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.50811 (* 0.0909091 = 0.22801 loss) | |
| I0506 01:31:05.967353 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.60796 (* 0.0909091 = 0.237087 loss) | |
| I0506 01:31:05.967366 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.05242 (* 0.0909091 = 0.186583 loss) | |
| I0506 01:31:05.967381 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.732637 (* 0.0909091 = 0.0666034 loss) | |
| I0506 01:31:05.967394 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.526012 (* 0.0909091 = 0.0478193 loss) | |
| I0506 01:31:05.967407 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.510588 (* 0.0909091 = 0.0464171 loss) | |
| I0506 01:31:05.967420 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.525153 (* 0.0909091 = 0.0477412 loss) | |
| I0506 01:31:05.967434 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.48706 (* 0.0909091 = 0.0442782 loss) | |
| I0506 01:31:05.967447 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0296824 (* 0.0909091 = 0.0026984 loss) | |
| I0506 01:31:05.967461 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0197288 (* 0.0909091 = 0.00179353 loss) | |
| I0506 01:31:05.967475 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0160996 (* 0.0909091 = 0.0014636 loss) | |
| I0506 01:31:05.967489 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0106057 (* 0.0909091 = 0.000964159 loss) | |
| I0506 01:31:05.967504 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00620007 (* 0.0909091 = 0.000563643 loss) | |
| I0506 01:31:05.967517 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00350373 (* 0.0909091 = 0.000318521 loss) | |
| I0506 01:31:05.967530 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00299023 (* 0.0909091 = 0.00027184 loss) | |
| I0506 01:31:05.967545 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00188039 (* 0.0909091 = 0.000170944 loss) | |
| I0506 01:31:05.967557 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00119178 (* 0.0909091 = 0.000108344 loss) | |
| I0506 01:31:05.967571 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000466759 (* 0.0909091 = 4.24326e-05 loss) | |
| I0506 01:31:05.967583 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:31:05.967594 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:31:05.967605 15760 solver.cpp:245] Train net output #149: total_confidence = 9.8025e-05 | |
| I0506 01:31:05.967625 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000730898 | |
| I0506 01:31:05.967639 15760 sgd_solver.cpp:106] Iteration 18500, lr = 0.001 | |
| I0506 01:31:26.194779 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 46.9968 > 30) by scale factor 0.638341 | |
| I0506 01:31:55.444499 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.7393 > 30) by scale factor 0.670551 | |
| I0506 01:32:53.348167 15760 solver.cpp:229] Iteration 19000, loss = 10.0253 | |
| I0506 01:32:53.348289 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.03125 | |
| I0506 01:32:53.348309 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:32:53.348322 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 01:32:53.348335 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:32:53.348346 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0506 01:32:53.348358 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 01:32:53.348369 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 01:32:53.348381 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0506 01:32:53.348393 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:32:53.348405 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 01:32:53.348417 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0506 01:32:53.348428 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:32:53.348440 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:32:53.348453 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 01:32:53.348464 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0506 01:32:53.348476 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0506 01:32:53.348489 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0506 01:32:53.348500 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0506 01:32:53.348512 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0506 01:32:53.348525 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0.875 | |
| I0506 01:32:53.348536 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 0.875 | |
| I0506 01:32:53.348548 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 0.875 | |
| I0506 01:32:53.348559 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:32:53.348572 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.647727 | |
| I0506 01:32:53.348583 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.28125 | |
| I0506 01:32:53.348608 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.27787 (* 0.3 = 0.98336 loss) | |
| I0506 01:32:53.348623 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.25076 (* 0.3 = 0.375228 loss) | |
| I0506 01:32:53.348637 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.3567 (* 0.0272727 = 0.0915463 loss) | |
| I0506 01:32:53.348650 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.58198 (* 0.0272727 = 0.0976903 loss) | |
| I0506 01:32:53.348664 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.13647 (* 0.0272727 = 0.08554 loss) | |
| I0506 01:32:53.348678 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.81663 (* 0.0272727 = 0.0768173 loss) | |
| I0506 01:32:53.348692 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.76021 (* 0.0272727 = 0.0752784 loss) | |
| I0506 01:32:53.348706 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.06974 (* 0.0272727 = 0.0837202 loss) | |
| I0506 01:32:53.348719 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.95295 (* 0.0272727 = 0.0532623 loss) | |
| I0506 01:32:53.348733 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.8757 (* 0.0272727 = 0.0238827 loss) | |
| I0506 01:32:53.348747 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 1.01478 (* 0.0272727 = 0.0276758 loss) | |
| I0506 01:32:53.348760 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 1.2079 (* 0.0272727 = 0.0329428 loss) | |
| I0506 01:32:53.348774 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.587191 (* 0.0272727 = 0.0160143 loss) | |
| I0506 01:32:53.348788 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.502086 (* 0.0272727 = 0.0136932 loss) | |
| I0506 01:32:53.348820 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.495775 (* 0.0272727 = 0.0135211 loss) | |
| I0506 01:32:53.348835 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.480941 (* 0.0272727 = 0.0131166 loss) | |
| I0506 01:32:53.348850 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.492761 (* 0.0272727 = 0.0134389 loss) | |
| I0506 01:32:53.348863 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.696535 (* 0.0272727 = 0.0189964 loss) | |
| I0506 01:32:53.348876 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.685813 (* 0.0272727 = 0.018704 loss) | |
| I0506 01:32:53.348891 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.769912 (* 0.0272727 = 0.0209976 loss) | |
| I0506 01:32:53.348904 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.78856 (* 0.0272727 = 0.0215062 loss) | |
| I0506 01:32:53.348917 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.836307 (* 0.0272727 = 0.0228084 loss) | |
| I0506 01:32:53.348932 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.80239 (* 0.0272727 = 0.0218834 loss) | |
| I0506 01:32:53.348947 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0181417 (* 0.0272727 = 0.000494774 loss) | |
| I0506 01:32:53.348961 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.046875 | |
| I0506 01:32:53.348974 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:32:53.348986 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:32:53.349001 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:32:53.349022 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:32:53.349036 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 01:32:53.349047 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 01:32:53.349059 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0506 01:32:53.349071 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:32:53.349083 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 01:32:53.349095 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0506 01:32:53.349103 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:32:53.349112 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:32:53.349138 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 01:32:53.349153 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0506 01:32:53.349164 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0506 01:32:53.349182 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0506 01:32:53.349195 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0506 01:32:53.349207 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0506 01:32:53.349220 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0.875 | |
| I0506 01:32:53.349231 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 0.875 | |
| I0506 01:32:53.349251 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 0.875 | |
| I0506 01:32:53.349266 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:32:53.349277 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.653409 | |
| I0506 01:32:53.349289 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.265625 | |
| I0506 01:32:53.349303 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.23589 (* 0.3 = 0.970767 loss) | |
| I0506 01:32:53.349318 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.20758 (* 0.3 = 0.362273 loss) | |
| I0506 01:32:53.349335 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.29988 (* 0.0272727 = 0.0899967 loss) | |
| I0506 01:32:53.349350 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.87484 (* 0.0272727 = 0.0784047 loss) | |
| I0506 01:32:53.349377 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 2.98521 (* 0.0272727 = 0.0814148 loss) | |
| I0506 01:32:53.349392 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.18563 (* 0.0272727 = 0.0868807 loss) | |
| I0506 01:32:53.349406 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.87786 (* 0.0272727 = 0.0784872 loss) | |
| I0506 01:32:53.349419 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.20318 (* 0.0272727 = 0.0873595 loss) | |
| I0506 01:32:53.349433 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.04121 (* 0.0272727 = 0.0556694 loss) | |
| I0506 01:32:53.349447 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.883892 (* 0.0272727 = 0.0241062 loss) | |
| I0506 01:32:53.349460 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 1.21122 (* 0.0272727 = 0.0330333 loss) | |
| I0506 01:32:53.349473 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 1.31744 (* 0.0272727 = 0.0359303 loss) | |
| I0506 01:32:53.349488 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.69748 (* 0.0272727 = 0.0190222 loss) | |
| I0506 01:32:53.349500 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.61417 (* 0.0272727 = 0.0167501 loss) | |
| I0506 01:32:53.349514 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.677156 (* 0.0272727 = 0.0184679 loss) | |
| I0506 01:32:53.349527 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.529993 (* 0.0272727 = 0.0144543 loss) | |
| I0506 01:32:53.349541 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.746355 (* 0.0272727 = 0.0203551 loss) | |
| I0506 01:32:53.349555 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.764952 (* 0.0272727 = 0.0208623 loss) | |
| I0506 01:32:53.349568 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.928908 (* 0.0272727 = 0.0253338 loss) | |
| I0506 01:32:53.349581 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.999314 (* 0.0272727 = 0.027254 loss) | |
| I0506 01:32:53.349594 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 1.0122 (* 0.0272727 = 0.0276056 loss) | |
| I0506 01:32:53.349608 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 1.15832 (* 0.0272727 = 0.0315906 loss) | |
| I0506 01:32:53.349622 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 1.1474 (* 0.0272727 = 0.0312928 loss) | |
| I0506 01:32:53.349635 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00194434 (* 0.0272727 = 5.30275e-05 loss) | |
| I0506 01:32:53.349647 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.078125 | |
| I0506 01:32:53.349658 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 01:32:53.349670 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0506 01:32:53.349681 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:32:53.349692 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:32:53.349704 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:32:53.349715 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:32:53.349726 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:32:53.349746 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:32:53.349766 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 01:32:53.349781 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0506 01:32:53.349792 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:32:53.349803 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:32:53.349815 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 01:32:53.349827 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0506 01:32:53.349838 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0506 01:32:53.349849 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0506 01:32:53.349871 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0506 01:32:53.349884 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0506 01:32:53.349896 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0.875 | |
| I0506 01:32:53.349907 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 0.875 | |
| I0506 01:32:53.349920 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 0.875 | |
| I0506 01:32:53.349931 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:32:53.349941 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.664773 | |
| I0506 01:32:53.349953 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.265625 | |
| I0506 01:32:53.349967 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.09378 (* 1 = 3.09378 loss) | |
| I0506 01:32:53.349980 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.14793 (* 1 = 1.14793 loss) | |
| I0506 01:32:53.349993 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.90204 (* 0.0909091 = 0.263822 loss) | |
| I0506 01:32:53.350010 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.77895 (* 0.0909091 = 0.252632 loss) | |
| I0506 01:32:53.350024 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.64331 (* 0.0909091 = 0.240301 loss) | |
| I0506 01:32:53.350039 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.65081 (* 0.0909091 = 0.240982 loss) | |
| I0506 01:32:53.350051 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.71967 (* 0.0909091 = 0.247243 loss) | |
| I0506 01:32:53.350064 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.87335 (* 0.0909091 = 0.261214 loss) | |
| I0506 01:32:53.350078 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.8812 (* 0.0909091 = 0.171018 loss) | |
| I0506 01:32:53.350091 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.657989 (* 0.0909091 = 0.0598171 loss) | |
| I0506 01:32:53.350105 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 1.06833 (* 0.0909091 = 0.0971209 loss) | |
| I0506 01:32:53.350118 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 1.14041 (* 0.0909091 = 0.103674 loss) | |
| I0506 01:32:53.350131 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.552255 (* 0.0909091 = 0.050205 loss) | |
| I0506 01:32:53.350145 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.46742 (* 0.0909091 = 0.0424927 loss) | |
| I0506 01:32:53.350158 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.453183 (* 0.0909091 = 0.0411985 loss) | |
| I0506 01:32:53.350172 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.418611 (* 0.0909091 = 0.0380556 loss) | |
| I0506 01:32:53.350186 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.610126 (* 0.0909091 = 0.055466 loss) | |
| I0506 01:32:53.350199 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.535275 (* 0.0909091 = 0.0486614 loss) | |
| I0506 01:32:53.350214 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.447987 (* 0.0909091 = 0.0407261 loss) | |
| I0506 01:32:53.350227 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.645515 (* 0.0909091 = 0.0586832 loss) | |
| I0506 01:32:53.350240 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.723813 (* 0.0909091 = 0.0658012 loss) | |
| I0506 01:32:53.350255 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.739833 (* 0.0909091 = 0.0672575 loss) | |
| I0506 01:32:53.350267 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.903198 (* 0.0909091 = 0.0821089 loss) | |
| I0506 01:32:53.350281 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00332593 (* 0.0909091 = 0.000302357 loss) | |
| I0506 01:32:53.350294 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:32:53.350306 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:32:53.350317 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000143586 | |
| I0506 01:32:53.350337 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000505558 | |
| I0506 01:32:53.350353 15760 sgd_solver.cpp:106] Iteration 19000, lr = 0.001 | |
| I0506 01:34:40.817276 15760 solver.cpp:229] Iteration 19500, loss = 9.90241 | |
| I0506 01:34:40.817433 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0816327 | |
| I0506 01:34:40.817453 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:34:40.817466 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:34:40.817479 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:34:40.817490 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 01:34:40.817502 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0506 01:34:40.817514 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0506 01:34:40.817525 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:34:40.817538 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 01:34:40.817549 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:34:40.817561 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:34:40.817572 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:34:40.817584 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:34:40.817595 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:34:40.817606 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:34:40.817618 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:34:40.817630 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:34:40.817641 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:34:40.817652 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:34:40.817663 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:34:40.817675 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:34:40.817687 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:34:40.817698 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:34:40.817708 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0506 01:34:40.817720 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.204082 | |
| I0506 01:34:40.817736 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.11162 (* 0.3 = 0.933486 loss) | |
| I0506 01:34:40.817750 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.981451 (* 0.3 = 0.294435 loss) | |
| I0506 01:34:40.817764 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.21496 (* 0.0272727 = 0.0876808 loss) | |
| I0506 01:34:40.817778 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.01258 (* 0.0272727 = 0.0821612 loss) | |
| I0506 01:34:40.817791 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.11137 (* 0.0272727 = 0.0848554 loss) | |
| I0506 01:34:40.817806 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.55032 (* 0.0272727 = 0.0968269 loss) | |
| I0506 01:34:40.817819 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.38588 (* 0.0272727 = 0.0923421 loss) | |
| I0506 01:34:40.817832 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.44171 (* 0.0272727 = 0.0938648 loss) | |
| I0506 01:34:40.817847 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.92601 (* 0.0272727 = 0.0525274 loss) | |
| I0506 01:34:40.817860 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.120298 (* 0.0272727 = 0.00328087 loss) | |
| I0506 01:34:40.817876 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0415266 (* 0.0272727 = 0.00113254 loss) | |
| I0506 01:34:40.817891 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0244901 (* 0.0272727 = 0.000667911 loss) | |
| I0506 01:34:40.817906 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0384318 (* 0.0272727 = 0.00104814 loss) | |
| I0506 01:34:40.817920 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0171975 (* 0.0272727 = 0.000469023 loss) | |
| I0506 01:34:40.817934 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0179786 (* 0.0272727 = 0.000490326 loss) | |
| I0506 01:34:40.817961 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0197122 (* 0.0272727 = 0.000537605 loss) | |
| I0506 01:34:40.817976 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0183153 (* 0.0272727 = 0.000499507 loss) | |
| I0506 01:34:40.817991 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0252694 (* 0.0272727 = 0.000689165 loss) | |
| I0506 01:34:40.818004 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0257167 (* 0.0272727 = 0.000701364 loss) | |
| I0506 01:34:40.818034 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.020058 (* 0.0272727 = 0.000547035 loss) | |
| I0506 01:34:40.818053 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0134699 (* 0.0272727 = 0.00036736 loss) | |
| I0506 01:34:40.818066 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0249889 (* 0.0272727 = 0.000681516 loss) | |
| I0506 01:34:40.818080 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0113371 (* 0.0272727 = 0.000309194 loss) | |
| I0506 01:34:40.818094 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0139992 (* 0.0272727 = 0.000381796 loss) | |
| I0506 01:34:40.818106 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0612245 | |
| I0506 01:34:40.818117 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:34:40.818130 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:34:40.818141 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:34:40.818152 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:34:40.818164 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 01:34:40.818176 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0506 01:34:40.818186 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:34:40.818198 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 01:34:40.818209 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:34:40.818220 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:34:40.818233 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:34:40.818243 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:34:40.818254 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:34:40.818265 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:34:40.818276 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:34:40.818287 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:34:40.818298 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:34:40.818310 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:34:40.818321 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:34:40.818332 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:34:40.818343 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:34:40.818354 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:34:40.818366 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.732955 | |
| I0506 01:34:40.818377 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.183673 | |
| I0506 01:34:40.818390 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.97176 (* 0.3 = 0.891527 loss) | |
| I0506 01:34:40.818404 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.910871 (* 0.3 = 0.273261 loss) | |
| I0506 01:34:40.818418 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.01163 (* 0.0272727 = 0.0821354 loss) | |
| I0506 01:34:40.818431 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.98883 (* 0.0272727 = 0.0815135 loss) | |
| I0506 01:34:40.818460 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.16004 (* 0.0272727 = 0.086183 loss) | |
| I0506 01:34:40.818475 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.12614 (* 0.0272727 = 0.0852582 loss) | |
| I0506 01:34:40.818490 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.08725 (* 0.0272727 = 0.0841976 loss) | |
| I0506 01:34:40.818503 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.36097 (* 0.0272727 = 0.0916628 loss) | |
| I0506 01:34:40.818516 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.78533 (* 0.0272727 = 0.0486909 loss) | |
| I0506 01:34:40.818531 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0909473 (* 0.0272727 = 0.00248038 loss) | |
| I0506 01:34:40.818543 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0314063 (* 0.0272727 = 0.000856534 loss) | |
| I0506 01:34:40.818557 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0179284 (* 0.0272727 = 0.000488956 loss) | |
| I0506 01:34:40.818572 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00566096 (* 0.0272727 = 0.00015439 loss) | |
| I0506 01:34:40.818584 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0110704 (* 0.0272727 = 0.00030192 loss) | |
| I0506 01:34:40.818598 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00606628 (* 0.0272727 = 0.000165444 loss) | |
| I0506 01:34:40.818613 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00645816 (* 0.0272727 = 0.000176132 loss) | |
| I0506 01:34:40.818625 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00504201 (* 0.0272727 = 0.000137509 loss) | |
| I0506 01:34:40.818639 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00348962 (* 0.0272727 = 9.51716e-05 loss) | |
| I0506 01:34:40.818653 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00537771 (* 0.0272727 = 0.000146665 loss) | |
| I0506 01:34:40.818667 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00289354 (* 0.0272727 = 7.89148e-05 loss) | |
| I0506 01:34:40.818681 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00461429 (* 0.0272727 = 0.000125844 loss) | |
| I0506 01:34:40.818694 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00402077 (* 0.0272727 = 0.000109657 loss) | |
| I0506 01:34:40.818708 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00359085 (* 0.0272727 = 9.79324e-05 loss) | |
| I0506 01:34:40.818722 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00202584 (* 0.0272727 = 5.52503e-05 loss) | |
| I0506 01:34:40.818733 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0816327 | |
| I0506 01:34:40.818745 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:34:40.818756 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:34:40.818768 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:34:40.818779 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:34:40.818790 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 01:34:40.818802 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 01:34:40.818814 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:34:40.818825 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 01:34:40.818835 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:34:40.818847 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:34:40.818858 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:34:40.818868 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:34:40.818879 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:34:40.818892 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:34:40.818902 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:34:40.818912 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:34:40.818936 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:34:40.818950 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:34:40.818961 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:34:40.818974 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:34:40.818984 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:34:40.818995 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:34:40.819006 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.738636 | |
| I0506 01:34:40.819018 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.285714 | |
| I0506 01:34:40.819031 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.84167 (* 1 = 2.84167 loss) | |
| I0506 01:34:40.819046 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.875722 (* 1 = 0.875722 loss) | |
| I0506 01:34:40.819059 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.71501 (* 0.0909091 = 0.246819 loss) | |
| I0506 01:34:40.819072 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.09289 (* 0.0909091 = 0.281172 loss) | |
| I0506 01:34:40.819087 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.10776 (* 0.0909091 = 0.282523 loss) | |
| I0506 01:34:40.819099 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.99168 (* 0.0909091 = 0.271971 loss) | |
| I0506 01:34:40.819113 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.36154 (* 0.0909091 = 0.214685 loss) | |
| I0506 01:34:40.819126 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.67615 (* 0.0909091 = 0.243286 loss) | |
| I0506 01:34:40.819139 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.44542 (* 0.0909091 = 0.131401 loss) | |
| I0506 01:34:40.819154 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.12107 (* 0.0909091 = 0.0110063 loss) | |
| I0506 01:34:40.819166 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0129517 (* 0.0909091 = 0.00117743 loss) | |
| I0506 01:34:40.819180 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00357616 (* 0.0909091 = 0.000325105 loss) | |
| I0506 01:34:40.819195 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00113624 (* 0.0909091 = 0.000103294 loss) | |
| I0506 01:34:40.819207 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00103679 (* 0.0909091 = 9.42536e-05 loss) | |
| I0506 01:34:40.819221 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00126625 (* 0.0909091 = 0.000115114 loss) | |
| I0506 01:34:40.819236 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000979683 (* 0.0909091 = 8.90621e-05 loss) | |
| I0506 01:34:40.819249 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000544621 (* 0.0909091 = 4.9511e-05 loss) | |
| I0506 01:34:40.819263 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000792618 (* 0.0909091 = 7.20562e-05 loss) | |
| I0506 01:34:40.819277 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000665278 (* 0.0909091 = 6.04798e-05 loss) | |
| I0506 01:34:40.819290 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000459807 (* 0.0909091 = 4.18006e-05 loss) | |
| I0506 01:34:40.819300 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000718965 (* 0.0909091 = 6.53605e-05 loss) | |
| I0506 01:34:40.819310 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000453162 (* 0.0909091 = 4.11965e-05 loss) | |
| I0506 01:34:40.819324 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000433362 (* 0.0909091 = 3.93965e-05 loss) | |
| I0506 01:34:40.819337 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000266776 (* 0.0909091 = 2.42523e-05 loss) | |
| I0506 01:34:40.819350 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:34:40.819360 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:34:40.819371 15760 solver.cpp:245] Train net output #149: total_confidence = 4.72565e-07 | |
| I0506 01:34:40.819392 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.43332e-06 | |
| I0506 01:34:40.819406 15760 sgd_solver.cpp:106] Iteration 19500, lr = 0.001 | |
| I0506 01:35:11.597625 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.4421 > 30) by scale factor 0.924723 | |
| I0506 01:36:27.951376 15760 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_20000.caffemodel | |
| I0506 01:36:28.399567 15760 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm15_bn_iter_20000.solverstate | |
| I0506 01:36:28.615901 15760 solver.cpp:338] Iteration 20000, Testing net (#0) | |
| I0506 01:37:04.665987 15760 solver.cpp:393] Test loss: 9.183 | |
| I0506 01:37:04.666117 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0663389 | |
| I0506 01:37:04.666137 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.124 | |
| I0506 01:37:04.666152 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.098 | |
| I0506 01:37:04.666163 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.079 | |
| I0506 01:37:04.666175 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.195 | |
| I0506 01:37:04.666187 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.299 | |
| I0506 01:37:04.666198 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.445 | |
| I0506 01:37:04.666209 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.726 | |
| I0506 01:37:04.666220 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.908 | |
| I0506 01:37:04.666232 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.99 | |
| I0506 01:37:04.666244 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.999 | |
| I0506 01:37:04.666254 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0506 01:37:04.666266 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0506 01:37:04.666277 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0506 01:37:04.666288 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0506 01:37:04.666301 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0506 01:37:04.666311 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0506 01:37:04.666323 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0506 01:37:04.666334 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0506 01:37:04.666345 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0506 01:37:04.666357 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0506 01:37:04.666368 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 01:37:04.666379 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 01:37:04.666390 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.76382 | |
| I0506 01:37:04.666402 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.230643 | |
| I0506 01:37:04.666425 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.58719 (* 0.3 = 1.07616 loss) | |
| I0506 01:37:04.666441 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.935042 (* 0.3 = 0.280513 loss) | |
| I0506 01:37:04.666455 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.01397 (* 0.0272727 = 0.0821991 loss) | |
| I0506 01:37:04.666468 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.13801 (* 0.0272727 = 0.0855821 loss) | |
| I0506 01:37:04.666482 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.25163 (* 0.0272727 = 0.0886809 loss) | |
| I0506 01:37:04.666496 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.03274 (* 0.0272727 = 0.082711 loss) | |
| I0506 01:37:04.666508 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 2.65562 (* 0.0272727 = 0.0724259 loss) | |
| I0506 01:37:04.666522 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.25847 (* 0.0272727 = 0.0615945 loss) | |
| I0506 01:37:04.666534 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.32846 (* 0.0272727 = 0.0362306 loss) | |
| I0506 01:37:04.666548 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 0.517447 (* 0.0272727 = 0.0141122 loss) | |
| I0506 01:37:04.666561 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.0766959 (* 0.0272727 = 0.00209171 loss) | |
| I0506 01:37:04.666575 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0337174 (* 0.0272727 = 0.000919565 loss) | |
| I0506 01:37:04.666589 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0238053 (* 0.0272727 = 0.000649237 loss) | |
| I0506 01:37:04.666602 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0172065 (* 0.0272727 = 0.000469268 loss) | |
| I0506 01:37:04.666616 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0133634 (* 0.0272727 = 0.000364456 loss) | |
| I0506 01:37:04.666651 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0106183 (* 0.0272727 = 0.000289589 loss) | |
| I0506 01:37:04.666666 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0077444 (* 0.0272727 = 0.000211211 loss) | |
| I0506 01:37:04.666679 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.00491199 (* 0.0272727 = 0.000133963 loss) | |
| I0506 01:37:04.666693 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.00294261 (* 0.0272727 = 8.02531e-05 loss) | |
| I0506 01:37:04.666707 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00268327 (* 0.0272727 = 7.318e-05 loss) | |
| I0506 01:37:04.666720 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00235571 (* 0.0272727 = 6.42466e-05 loss) | |
| I0506 01:37:04.666733 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00211555 (* 0.0272727 = 5.76969e-05 loss) | |
| I0506 01:37:04.666748 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00202896 (* 0.0272727 = 5.53351e-05 loss) | |
| I0506 01:37:04.666760 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00195844 (* 0.0272727 = 5.34119e-05 loss) | |
| I0506 01:37:04.666772 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0679004 | |
| I0506 01:37:04.666784 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.132 | |
| I0506 01:37:04.666795 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.095 | |
| I0506 01:37:04.666806 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.092 | |
| I0506 01:37:04.666817 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.183 | |
| I0506 01:37:04.666828 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.317 | |
| I0506 01:37:04.666841 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.441 | |
| I0506 01:37:04.666851 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.724 | |
| I0506 01:37:04.666862 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.907 | |
| I0506 01:37:04.666877 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.99 | |
| I0506 01:37:04.666888 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.999 | |
| I0506 01:37:04.666900 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0506 01:37:04.666911 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0506 01:37:04.666923 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0506 01:37:04.666934 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0506 01:37:04.666945 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0506 01:37:04.666952 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0506 01:37:04.666960 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0506 01:37:04.666971 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0506 01:37:04.666982 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0506 01:37:04.666993 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0506 01:37:04.667004 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 01:37:04.667016 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 01:37:04.667032 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.763729 | |
| I0506 01:37:04.667045 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.234064 | |
| I0506 01:37:04.667058 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.5501 (* 0.3 = 1.06503 loss) | |
| I0506 01:37:04.667073 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.925855 (* 0.3 = 0.277757 loss) | |
| I0506 01:37:04.667086 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 2.96445 (* 0.0272727 = 0.0808488 loss) | |
| I0506 01:37:04.667099 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.08507 (* 0.0272727 = 0.0841383 loss) | |
| I0506 01:37:04.667112 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.20373 (* 0.0272727 = 0.0873745 loss) | |
| I0506 01:37:04.667142 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 2.98155 (* 0.0272727 = 0.081315 loss) | |
| I0506 01:37:04.667157 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 2.6168 (* 0.0272727 = 0.0713671 loss) | |
| I0506 01:37:04.667171 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.22296 (* 0.0272727 = 0.0606262 loss) | |
| I0506 01:37:04.667184 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.30553 (* 0.0272727 = 0.0356053 loss) | |
| I0506 01:37:04.667197 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 0.513095 (* 0.0272727 = 0.0139935 loss) | |
| I0506 01:37:04.667210 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.0788435 (* 0.0272727 = 0.00215028 loss) | |
| I0506 01:37:04.667224 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0345731 (* 0.0272727 = 0.000942903 loss) | |
| I0506 01:37:04.667237 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0245209 (* 0.0272727 = 0.000668751 loss) | |
| I0506 01:37:04.667250 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0178599 (* 0.0272727 = 0.000487087 loss) | |
| I0506 01:37:04.667264 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0157482 (* 0.0272727 = 0.000429497 loss) | |
| I0506 01:37:04.667278 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.011043 (* 0.0272727 = 0.000301174 loss) | |
| I0506 01:37:04.667290 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00894737 (* 0.0272727 = 0.000244019 loss) | |
| I0506 01:37:04.667304 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00562217 (* 0.0272727 = 0.000153332 loss) | |
| I0506 01:37:04.667316 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.00353906 (* 0.0272727 = 9.65197e-05 loss) | |
| I0506 01:37:04.667330 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00295099 (* 0.0272727 = 8.04815e-05 loss) | |
| I0506 01:37:04.667343 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.00247464 (* 0.0272727 = 6.74903e-05 loss) | |
| I0506 01:37:04.667356 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00242008 (* 0.0272727 = 6.60022e-05 loss) | |
| I0506 01:37:04.667369 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00220097 (* 0.0272727 = 6.00264e-05 loss) | |
| I0506 01:37:04.667382 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00220788 (* 0.0272727 = 6.0215e-05 loss) | |
| I0506 01:37:04.667393 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.101381 | |
| I0506 01:37:04.667405 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.145 | |
| I0506 01:37:04.667417 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.117 | |
| I0506 01:37:04.667428 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.087 | |
| I0506 01:37:04.667438 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.192 | |
| I0506 01:37:04.667449 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.327 | |
| I0506 01:37:04.667459 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.443 | |
| I0506 01:37:04.667471 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.723 | |
| I0506 01:37:04.667481 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.904 | |
| I0506 01:37:04.667493 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.988 | |
| I0506 01:37:04.667505 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.997 | |
| I0506 01:37:04.667515 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.999 | |
| I0506 01:37:04.667526 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.999 | |
| I0506 01:37:04.667537 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.999 | |
| I0506 01:37:04.667548 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.999 | |
| I0506 01:37:04.667559 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0506 01:37:04.667569 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0506 01:37:04.667592 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0506 01:37:04.667603 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0506 01:37:04.667614 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0506 01:37:04.667625 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0506 01:37:04.667635 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 01:37:04.667646 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 01:37:04.667657 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.766502 | |
| I0506 01:37:04.667668 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.267058 | |
| I0506 01:37:04.667682 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 2.98996 (* 1 = 2.98996 loss) | |
| I0506 01:37:04.667695 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.822966 (* 1 = 0.822966 loss) | |
| I0506 01:37:04.667708 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 2.7696 (* 0.0909091 = 0.251782 loss) | |
| I0506 01:37:04.667721 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 2.90462 (* 0.0909091 = 0.264056 loss) | |
| I0506 01:37:04.667734 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.03938 (* 0.0909091 = 0.276307 loss) | |
| I0506 01:37:04.667747 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 2.80102 (* 0.0909091 = 0.254638 loss) | |
| I0506 01:37:04.667760 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 2.41963 (* 0.0909091 = 0.219966 loss) | |
| I0506 01:37:04.667773 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.06338 (* 0.0909091 = 0.18758 loss) | |
| I0506 01:37:04.667786 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.1725 (* 0.0909091 = 0.106591 loss) | |
| I0506 01:37:04.667799 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.456325 (* 0.0909091 = 0.0414841 loss) | |
| I0506 01:37:04.667812 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.0789594 (* 0.0909091 = 0.00717813 loss) | |
| I0506 01:37:04.667826 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0324539 (* 0.0909091 = 0.00295036 loss) | |
| I0506 01:37:04.667840 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0220152 (* 0.0909091 = 0.00200139 loss) | |
| I0506 01:37:04.667852 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0159475 (* 0.0909091 = 0.00144977 loss) | |
| I0506 01:37:04.667865 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0126234 (* 0.0909091 = 0.00114758 loss) | |
| I0506 01:37:04.667878 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.00924947 (* 0.0909091 = 0.000840861 loss) | |
| I0506 01:37:04.667892 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.00703806 (* 0.0909091 = 0.000639824 loss) | |
| I0506 01:37:04.667906 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00516888 (* 0.0909091 = 0.000469898 loss) | |
| I0506 01:37:04.667918 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00292683 (* 0.0909091 = 0.000266075 loss) | |
| I0506 01:37:04.667935 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00307141 (* 0.0909091 = 0.000279219 loss) | |
| I0506 01:37:04.667949 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00288985 (* 0.0909091 = 0.000262714 loss) | |
| I0506 01:37:04.667963 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0024279 (* 0.0909091 = 0.000220718 loss) | |
| I0506 01:37:04.667976 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00208389 (* 0.0909091 = 0.000189445 loss) | |
| I0506 01:37:04.667989 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00207489 (* 0.0909091 = 0.000188626 loss) | |
| I0506 01:37:04.668000 15760 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0506 01:37:04.668011 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:37:04.668022 15760 solver.cpp:406] Test net output #149: total_confidence = 0.00032937 | |
| I0506 01:37:04.668043 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000532796 | |
| I0506 01:37:04.668057 15760 solver.cpp:338] Iteration 20000, Testing net (#1) | |
| I0506 01:37:40.583534 15760 solver.cpp:393] Test loss: 9.78443 | |
| I0506 01:37:40.583657 15760 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0609514 | |
| I0506 01:37:40.583678 15760 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.124 | |
| I0506 01:37:40.583690 15760 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.094 | |
| I0506 01:37:40.583703 15760 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.098 | |
| I0506 01:37:40.583714 15760 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.184 | |
| I0506 01:37:40.583726 15760 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.31 | |
| I0506 01:37:40.583739 15760 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.423 | |
| I0506 01:37:40.583750 15760 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.634 | |
| I0506 01:37:40.583761 15760 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.799 | |
| I0506 01:37:40.583772 15760 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.886 | |
| I0506 01:37:40.583784 15760 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.907 | |
| I0506 01:37:40.583796 15760 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.923 | |
| I0506 01:37:40.583806 15760 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.936 | |
| I0506 01:37:40.583818 15760 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.948 | |
| I0506 01:37:40.583829 15760 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.96 | |
| I0506 01:37:40.583840 15760 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.97 | |
| I0506 01:37:40.583853 15760 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.979 | |
| I0506 01:37:40.583863 15760 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.989 | |
| I0506 01:37:40.583878 15760 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.996 | |
| I0506 01:37:40.583890 15760 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.997 | |
| I0506 01:37:40.583901 15760 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.999 | |
| I0506 01:37:40.583912 15760 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0506 01:37:40.583925 15760 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0506 01:37:40.583936 15760 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.729364 | |
| I0506 01:37:40.583947 15760 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.234498 | |
| I0506 01:37:40.583963 15760 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.57385 (* 0.3 = 1.07216 loss) | |
| I0506 01:37:40.583977 15760 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.0628 (* 0.3 = 0.318841 loss) | |
| I0506 01:37:40.583992 15760 solver.cpp:406] Test net output #27: loss1/loss01 = 3.01882 (* 0.0272727 = 0.0823315 loss) | |
| I0506 01:37:40.584004 15760 solver.cpp:406] Test net output #28: loss1/loss02 = 3.14569 (* 0.0272727 = 0.0857915 loss) | |
| I0506 01:37:40.584018 15760 solver.cpp:406] Test net output #29: loss1/loss03 = 3.21529 (* 0.0272727 = 0.0876897 loss) | |
| I0506 01:37:40.584031 15760 solver.cpp:406] Test net output #30: loss1/loss04 = 3.06394 (* 0.0272727 = 0.083562 loss) | |
| I0506 01:37:40.584044 15760 solver.cpp:406] Test net output #31: loss1/loss05 = 2.6697 (* 0.0272727 = 0.07281 loss) | |
| I0506 01:37:40.584059 15760 solver.cpp:406] Test net output #32: loss1/loss06 = 2.34564 (* 0.0272727 = 0.063972 loss) | |
| I0506 01:37:40.584071 15760 solver.cpp:406] Test net output #33: loss1/loss07 = 1.59404 (* 0.0272727 = 0.0434737 loss) | |
| I0506 01:37:40.584084 15760 solver.cpp:406] Test net output #34: loss1/loss08 = 0.913763 (* 0.0272727 = 0.0249208 loss) | |
| I0506 01:37:40.584098 15760 solver.cpp:406] Test net output #35: loss1/loss09 = 0.488729 (* 0.0272727 = 0.013329 loss) | |
| I0506 01:37:40.584111 15760 solver.cpp:406] Test net output #36: loss1/loss10 = 0.397615 (* 0.0272727 = 0.010844 loss) | |
| I0506 01:37:40.584125 15760 solver.cpp:406] Test net output #37: loss1/loss11 = 0.339222 (* 0.0272727 = 0.00925151 loss) | |
| I0506 01:37:40.584138 15760 solver.cpp:406] Test net output #38: loss1/loss12 = 0.30214 (* 0.0272727 = 0.00824018 loss) | |
| I0506 01:37:40.584152 15760 solver.cpp:406] Test net output #39: loss1/loss13 = 0.259943 (* 0.0272727 = 0.00708934 loss) | |
| I0506 01:37:40.586251 15760 solver.cpp:406] Test net output #40: loss1/loss14 = 0.213387 (* 0.0272727 = 0.00581964 loss) | |
| I0506 01:37:40.586267 15760 solver.cpp:406] Test net output #41: loss1/loss15 = 0.165606 (* 0.0272727 = 0.00451652 loss) | |
| I0506 01:37:40.586282 15760 solver.cpp:406] Test net output #42: loss1/loss16 = 0.12858 (* 0.0272727 = 0.00350674 loss) | |
| I0506 01:37:40.586294 15760 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0814644 (* 0.0272727 = 0.00222176 loss) | |
| I0506 01:37:40.586308 15760 solver.cpp:406] Test net output #44: loss1/loss18 = 0.035316 (* 0.0272727 = 0.000963163 loss) | |
| I0506 01:37:40.586323 15760 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0273294 (* 0.0272727 = 0.000745347 loss) | |
| I0506 01:37:40.586336 15760 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0102829 (* 0.0272727 = 0.000280442 loss) | |
| I0506 01:37:40.586349 15760 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00247842 (* 0.0272727 = 6.75932e-05 loss) | |
| I0506 01:37:40.586364 15760 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00225595 (* 0.0272727 = 6.15258e-05 loss) | |
| I0506 01:37:40.586375 15760 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0690248 | |
| I0506 01:37:40.586386 15760 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.123 | |
| I0506 01:37:40.586398 15760 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.12 | |
| I0506 01:37:40.586410 15760 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.107 | |
| I0506 01:37:40.586421 15760 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.18 | |
| I0506 01:37:40.586432 15760 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.311 | |
| I0506 01:37:40.586443 15760 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.418 | |
| I0506 01:37:40.586454 15760 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.634 | |
| I0506 01:37:40.586467 15760 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.8 | |
| I0506 01:37:40.586477 15760 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.887 | |
| I0506 01:37:40.586488 15760 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.907 | |
| I0506 01:37:40.586499 15760 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.923 | |
| I0506 01:37:40.586510 15760 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.936 | |
| I0506 01:37:40.586521 15760 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.948 | |
| I0506 01:37:40.586534 15760 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.96 | |
| I0506 01:37:40.586544 15760 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.97 | |
| I0506 01:37:40.586556 15760 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.979 | |
| I0506 01:37:40.586567 15760 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.989 | |
| I0506 01:37:40.586578 15760 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.996 | |
| I0506 01:37:40.586591 15760 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.997 | |
| I0506 01:37:40.586601 15760 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.999 | |
| I0506 01:37:40.586612 15760 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0506 01:37:40.586623 15760 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0506 01:37:40.586634 15760 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.731501 | |
| I0506 01:37:40.586645 15760 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.23616 | |
| I0506 01:37:40.586659 15760 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.5317 (* 0.3 = 1.05951 loss) | |
| I0506 01:37:40.586673 15760 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.04604 (* 0.3 = 0.313813 loss) | |
| I0506 01:37:40.586685 15760 solver.cpp:406] Test net output #76: loss2/loss01 = 2.97754 (* 0.0272727 = 0.0812055 loss) | |
| I0506 01:37:40.586699 15760 solver.cpp:406] Test net output #77: loss2/loss02 = 3.10165 (* 0.0272727 = 0.0845905 loss) | |
| I0506 01:37:40.586725 15760 solver.cpp:406] Test net output #78: loss2/loss03 = 3.15782 (* 0.0272727 = 0.0861224 loss) | |
| I0506 01:37:40.586740 15760 solver.cpp:406] Test net output #79: loss2/loss04 = 3.01433 (* 0.0272727 = 0.0822089 loss) | |
| I0506 01:37:40.586766 15760 solver.cpp:406] Test net output #80: loss2/loss05 = 2.63497 (* 0.0272727 = 0.0718628 loss) | |
| I0506 01:37:40.586783 15760 solver.cpp:406] Test net output #81: loss2/loss06 = 2.32954 (* 0.0272727 = 0.0635329 loss) | |
| I0506 01:37:40.586796 15760 solver.cpp:406] Test net output #82: loss2/loss07 = 1.58228 (* 0.0272727 = 0.043153 loss) | |
| I0506 01:37:40.586810 15760 solver.cpp:406] Test net output #83: loss2/loss08 = 0.897453 (* 0.0272727 = 0.024476 loss) | |
| I0506 01:37:40.586823 15760 solver.cpp:406] Test net output #84: loss2/loss09 = 0.48924 (* 0.0272727 = 0.0133429 loss) | |
| I0506 01:37:40.586838 15760 solver.cpp:406] Test net output #85: loss2/loss10 = 0.386867 (* 0.0272727 = 0.0105509 loss) | |
| I0506 01:37:40.586850 15760 solver.cpp:406] Test net output #86: loss2/loss11 = 0.332958 (* 0.0272727 = 0.00908067 loss) | |
| I0506 01:37:40.586864 15760 solver.cpp:406] Test net output #87: loss2/loss12 = 0.300982 (* 0.0272727 = 0.0082086 loss) | |
| I0506 01:37:40.586881 15760 solver.cpp:406] Test net output #88: loss2/loss13 = 0.250874 (* 0.0272727 = 0.00684201 loss) | |
| I0506 01:37:40.586895 15760 solver.cpp:406] Test net output #89: loss2/loss14 = 0.21091 (* 0.0272727 = 0.00575209 loss) | |
| I0506 01:37:40.586908 15760 solver.cpp:406] Test net output #90: loss2/loss15 = 0.163571 (* 0.0272727 = 0.00446102 loss) | |
| I0506 01:37:40.586922 15760 solver.cpp:406] Test net output #91: loss2/loss16 = 0.129346 (* 0.0272727 = 0.00352763 loss) | |
| I0506 01:37:40.586935 15760 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0790065 (* 0.0272727 = 0.00215472 loss) | |
| I0506 01:37:40.586949 15760 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0324925 (* 0.0272727 = 0.000886159 loss) | |
| I0506 01:37:40.586962 15760 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0262327 (* 0.0272727 = 0.000715438 loss) | |
| I0506 01:37:40.586976 15760 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0116936 (* 0.0272727 = 0.000318916 loss) | |
| I0506 01:37:40.586989 15760 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00318833 (* 0.0272727 = 8.69546e-05 loss) | |
| I0506 01:37:40.587002 15760 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00293142 (* 0.0272727 = 7.99478e-05 loss) | |
| I0506 01:37:40.587014 15760 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0987928 | |
| I0506 01:37:40.587025 15760 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.137 | |
| I0506 01:37:40.587038 15760 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.107 | |
| I0506 01:37:40.587049 15760 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.099 | |
| I0506 01:37:40.587059 15760 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.178 | |
| I0506 01:37:40.587071 15760 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.335 | |
| I0506 01:37:40.587081 15760 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.434 | |
| I0506 01:37:40.587093 15760 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.635 | |
| I0506 01:37:40.587105 15760 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.807 | |
| I0506 01:37:40.587116 15760 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.893 | |
| I0506 01:37:40.587126 15760 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.912 | |
| I0506 01:37:40.587137 15760 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.925 | |
| I0506 01:37:40.587148 15760 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.935 | |
| I0506 01:37:40.587159 15760 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.949 | |
| I0506 01:37:40.587170 15760 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.96 | |
| I0506 01:37:40.587182 15760 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.97 | |
| I0506 01:37:40.587193 15760 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.979 | |
| I0506 01:37:40.587214 15760 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.989 | |
| I0506 01:37:40.587227 15760 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.996 | |
| I0506 01:37:40.587239 15760 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.997 | |
| I0506 01:37:40.587255 15760 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.999 | |
| I0506 01:37:40.587266 15760 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0506 01:37:40.587278 15760 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0506 01:37:40.587286 15760 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.735364 | |
| I0506 01:37:40.587297 15760 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.267207 | |
| I0506 01:37:40.587311 15760 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.00465 (* 1 = 3.00465 loss) | |
| I0506 01:37:40.587324 15760 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.940135 (* 1 = 0.940135 loss) | |
| I0506 01:37:40.587338 15760 solver.cpp:406] Test net output #125: loss3/loss01 = 2.78026 (* 0.0909091 = 0.25275 loss) | |
| I0506 01:37:40.587352 15760 solver.cpp:406] Test net output #126: loss3/loss02 = 2.93738 (* 0.0909091 = 0.267035 loss) | |
| I0506 01:37:40.587364 15760 solver.cpp:406] Test net output #127: loss3/loss03 = 3.00608 (* 0.0909091 = 0.27328 loss) | |
| I0506 01:37:40.587378 15760 solver.cpp:406] Test net output #128: loss3/loss04 = 2.839 (* 0.0909091 = 0.258091 loss) | |
| I0506 01:37:40.587391 15760 solver.cpp:406] Test net output #129: loss3/loss05 = 2.42774 (* 0.0909091 = 0.220704 loss) | |
| I0506 01:37:40.587404 15760 solver.cpp:406] Test net output #130: loss3/loss06 = 2.14946 (* 0.0909091 = 0.195405 loss) | |
| I0506 01:37:40.587417 15760 solver.cpp:406] Test net output #131: loss3/loss07 = 1.42402 (* 0.0909091 = 0.129457 loss) | |
| I0506 01:37:40.587431 15760 solver.cpp:406] Test net output #132: loss3/loss08 = 0.799061 (* 0.0909091 = 0.0726419 loss) | |
| I0506 01:37:40.587445 15760 solver.cpp:406] Test net output #133: loss3/loss09 = 0.436925 (* 0.0909091 = 0.0397205 loss) | |
| I0506 01:37:40.587458 15760 solver.cpp:406] Test net output #134: loss3/loss10 = 0.341707 (* 0.0909091 = 0.0310643 loss) | |
| I0506 01:37:40.587471 15760 solver.cpp:406] Test net output #135: loss3/loss11 = 0.289849 (* 0.0909091 = 0.0263499 loss) | |
| I0506 01:37:40.587486 15760 solver.cpp:406] Test net output #136: loss3/loss12 = 0.266267 (* 0.0909091 = 0.0242061 loss) | |
| I0506 01:37:40.587498 15760 solver.cpp:406] Test net output #137: loss3/loss13 = 0.219769 (* 0.0909091 = 0.019979 loss) | |
| I0506 01:37:40.587512 15760 solver.cpp:406] Test net output #138: loss3/loss14 = 0.175977 (* 0.0909091 = 0.0159979 loss) | |
| I0506 01:37:40.587525 15760 solver.cpp:406] Test net output #139: loss3/loss15 = 0.136791 (* 0.0909091 = 0.0124355 loss) | |
| I0506 01:37:40.587538 15760 solver.cpp:406] Test net output #140: loss3/loss16 = 0.107394 (* 0.0909091 = 0.0097631 loss) | |
| I0506 01:37:40.587551 15760 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0640844 (* 0.0909091 = 0.00582585 loss) | |
| I0506 01:37:40.587565 15760 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0305451 (* 0.0909091 = 0.00277683 loss) | |
| I0506 01:37:40.587579 15760 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0229486 (* 0.0909091 = 0.00208624 loss) | |
| I0506 01:37:40.587591 15760 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00961045 (* 0.0909091 = 0.000873677 loss) | |
| I0506 01:37:40.587604 15760 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00154461 (* 0.0909091 = 0.000140419 loss) | |
| I0506 01:37:40.587618 15760 solver.cpp:406] Test net output #146: loss3/loss22 = 0.000991224 (* 0.0909091 = 9.01113e-05 loss) | |
| I0506 01:37:40.587630 15760 solver.cpp:406] Test net output #147: total_accuracy = 0.001 | |
| I0506 01:37:40.587641 15760 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0.002 | |
| I0506 01:37:40.587651 15760 solver.cpp:406] Test net output #149: total_confidence = 0.000329749 | |
| I0506 01:37:40.587672 15760 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000419286 | |
| I0506 01:37:40.722921 15760 solver.cpp:229] Iteration 20000, loss = 9.77093 | |
| I0506 01:37:40.722982 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0212766 | |
| I0506 01:37:40.723001 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:37:40.723013 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 01:37:40.723026 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0506 01:37:40.723037 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:37:40.723049 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:37:40.723062 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0506 01:37:40.723073 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0506 01:37:40.723085 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:37:40.723098 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:37:40.723109 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:37:40.723121 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:37:40.723132 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:37:40.723145 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:37:40.723157 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:37:40.723170 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:37:40.723181 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:37:40.723193 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:37:40.723204 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:37:40.723217 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:37:40.723228 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:37:40.723239 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:37:40.723251 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:37:40.723263 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0506 01:37:40.723274 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.12766 | |
| I0506 01:37:40.723289 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.15 (* 0.3 = 0.945 loss) | |
| I0506 01:37:40.723304 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.893099 (* 0.3 = 0.26793 loss) | |
| I0506 01:37:40.723317 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.84196 (* 0.0272727 = 0.0775079 loss) | |
| I0506 01:37:40.723331 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.2981 (* 0.0272727 = 0.0899481 loss) | |
| I0506 01:37:40.723345 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 2.91562 (* 0.0272727 = 0.079517 loss) | |
| I0506 01:37:40.723359 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.97304 (* 0.0272727 = 0.0810829 loss) | |
| I0506 01:37:40.723373 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.90835 (* 0.0272727 = 0.0793187 loss) | |
| I0506 01:37:40.723387 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.86407 (* 0.0272727 = 0.078111 loss) | |
| I0506 01:37:40.723400 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.79675 (* 0.0272727 = 0.0490024 loss) | |
| I0506 01:37:40.723414 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.55213 (* 0.0272727 = 0.0423307 loss) | |
| I0506 01:37:40.723428 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.078158 (* 0.0272727 = 0.00213158 loss) | |
| I0506 01:37:40.723443 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0585067 (* 0.0272727 = 0.00159564 loss) | |
| I0506 01:37:40.723458 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0310441 (* 0.0272727 = 0.000846657 loss) | |
| I0506 01:37:40.723497 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0197639 (* 0.0272727 = 0.000539016 loss) | |
| I0506 01:37:40.723512 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0192038 (* 0.0272727 = 0.000523741 loss) | |
| I0506 01:37:40.723526 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0135224 (* 0.0272727 = 0.000368793 loss) | |
| I0506 01:37:40.723546 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0134936 (* 0.0272727 = 0.000368008 loss) | |
| I0506 01:37:40.723559 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00926495 (* 0.0272727 = 0.00025268 loss) | |
| I0506 01:37:40.723573 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00451709 (* 0.0272727 = 0.000123193 loss) | |
| I0506 01:37:40.723587 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00634533 (* 0.0272727 = 0.000173054 loss) | |
| I0506 01:37:40.723601 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00415965 (* 0.0272727 = 0.000113445 loss) | |
| I0506 01:37:40.723615 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00547367 (* 0.0272727 = 0.000149282 loss) | |
| I0506 01:37:40.723629 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00699848 (* 0.0272727 = 0.000190868 loss) | |
| I0506 01:37:40.723644 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00353863 (* 0.0272727 = 9.6508e-05 loss) | |
| I0506 01:37:40.723655 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0638298 | |
| I0506 01:37:40.723669 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:37:40.723680 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:37:40.723692 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:37:40.723703 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:37:40.723716 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 01:37:40.723726 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0506 01:37:40.723738 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0506 01:37:40.723750 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:37:40.723762 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:37:40.723773 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:37:40.723785 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:37:40.723796 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:37:40.723808 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:37:40.723819 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:37:40.723830 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:37:40.723841 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:37:40.723853 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:37:40.723865 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:37:40.723875 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:37:40.723887 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:37:40.723898 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:37:40.723911 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:37:40.723922 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0506 01:37:40.723933 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.148936 | |
| I0506 01:37:40.723950 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.20816 (* 0.3 = 0.962447 loss) | |
| I0506 01:37:40.723965 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.943109 (* 0.3 = 0.282933 loss) | |
| I0506 01:37:40.723979 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.1195 (* 0.0272727 = 0.0850774 loss) | |
| I0506 01:37:40.724004 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.47104 (* 0.0272727 = 0.0946647 loss) | |
| I0506 01:37:40.724019 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.27869 (* 0.0272727 = 0.0894189 loss) | |
| I0506 01:37:40.724032 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 2.9131 (* 0.0272727 = 0.0794481 loss) | |
| I0506 01:37:40.724045 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.36797 (* 0.0272727 = 0.0918537 loss) | |
| I0506 01:37:40.724059 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.26908 (* 0.0272727 = 0.0618839 loss) | |
| I0506 01:37:40.724073 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.13989 (* 0.0272727 = 0.0583606 loss) | |
| I0506 01:37:40.724086 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.65436 (* 0.0272727 = 0.0451189 loss) | |
| I0506 01:37:40.724100 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.139 (* 0.0272727 = 0.00379091 loss) | |
| I0506 01:37:40.724114 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0911022 (* 0.0272727 = 0.00248461 loss) | |
| I0506 01:37:40.724128 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0705299 (* 0.0272727 = 0.00192354 loss) | |
| I0506 01:37:40.724143 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0440521 (* 0.0272727 = 0.00120142 loss) | |
| I0506 01:37:40.724156 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0403902 (* 0.0272727 = 0.00110155 loss) | |
| I0506 01:37:40.724169 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0318146 (* 0.0272727 = 0.000867672 loss) | |
| I0506 01:37:40.724184 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0207052 (* 0.0272727 = 0.000564689 loss) | |
| I0506 01:37:40.724196 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0228364 (* 0.0272727 = 0.00062281 loss) | |
| I0506 01:37:40.724210 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00891846 (* 0.0272727 = 0.000243231 loss) | |
| I0506 01:37:40.724225 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0092248 (* 0.0272727 = 0.000251586 loss) | |
| I0506 01:37:40.724237 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00521517 (* 0.0272727 = 0.000142232 loss) | |
| I0506 01:37:40.724251 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00449229 (* 0.0272727 = 0.000122517 loss) | |
| I0506 01:37:40.724266 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00599615 (* 0.0272727 = 0.000163531 loss) | |
| I0506 01:37:40.724279 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00461278 (* 0.0272727 = 0.000125803 loss) | |
| I0506 01:37:40.724292 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0851064 | |
| I0506 01:37:40.724303 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:37:40.724315 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 01:37:40.724326 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:37:40.724337 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:37:40.724349 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:37:40.724360 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0506 01:37:40.724372 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0506 01:37:40.724385 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:37:40.724396 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:37:40.724407 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:37:40.724418 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:37:40.724431 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:37:40.724442 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:37:40.724453 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:37:40.724473 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:37:40.724486 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:37:40.724498 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:37:40.724509 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:37:40.724520 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:37:40.724531 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:37:40.724542 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:37:40.724555 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:37:40.724565 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0506 01:37:40.724577 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.191489 | |
| I0506 01:37:40.724594 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.96439 (* 1 = 2.96439 loss) | |
| I0506 01:37:40.724609 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.870427 (* 1 = 0.870427 loss) | |
| I0506 01:37:40.724622 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.69713 (* 0.0909091 = 0.245194 loss) | |
| I0506 01:37:40.724637 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.18306 (* 0.0909091 = 0.289369 loss) | |
| I0506 01:37:40.724650 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.90056 (* 0.0909091 = 0.263687 loss) | |
| I0506 01:37:40.724663 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 2.73037 (* 0.0909091 = 0.248216 loss) | |
| I0506 01:37:40.724678 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.45187 (* 0.0909091 = 0.222897 loss) | |
| I0506 01:37:40.724690 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.13643 (* 0.0909091 = 0.19422 loss) | |
| I0506 01:37:40.724704 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.69823 (* 0.0909091 = 0.154384 loss) | |
| I0506 01:37:40.724717 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.34639 (* 0.0909091 = 0.122399 loss) | |
| I0506 01:37:40.724731 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0389435 (* 0.0909091 = 0.00354032 loss) | |
| I0506 01:37:40.724745 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0232874 (* 0.0909091 = 0.00211704 loss) | |
| I0506 01:37:40.724758 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0105036 (* 0.0909091 = 0.00095487 loss) | |
| I0506 01:37:40.724772 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00579453 (* 0.0909091 = 0.000526776 loss) | |
| I0506 01:37:40.724786 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00510447 (* 0.0909091 = 0.000464042 loss) | |
| I0506 01:37:40.724799 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00389442 (* 0.0909091 = 0.000354038 loss) | |
| I0506 01:37:40.724813 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00294994 (* 0.0909091 = 0.000268176 loss) | |
| I0506 01:37:40.724828 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00285326 (* 0.0909091 = 0.000259387 loss) | |
| I0506 01:37:40.724840 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00229571 (* 0.0909091 = 0.000208701 loss) | |
| I0506 01:37:40.724854 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00166039 (* 0.0909091 = 0.000150945 loss) | |
| I0506 01:37:40.724867 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00159168 (* 0.0909091 = 0.000144699 loss) | |
| I0506 01:37:40.724881 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000832416 (* 0.0909091 = 7.56742e-05 loss) | |
| I0506 01:37:40.724895 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000467003 (* 0.0909091 = 4.24548e-05 loss) | |
| I0506 01:37:40.724910 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000140476 (* 0.0909091 = 1.27705e-05 loss) | |
| I0506 01:37:40.724921 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:37:40.724942 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:37:40.724954 15760 solver.cpp:245] Train net output #149: total_confidence = 4.80083e-05 | |
| I0506 01:37:40.724967 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000170264 | |
| I0506 01:37:40.724979 15760 sgd_solver.cpp:106] Iteration 20000, lr = 0.001 | |
| I0506 01:37:52.347738 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.0424 > 30) by scale factor 0.730951 | |
| I0506 01:38:12.056915 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 52.3044 > 30) by scale factor 0.573566 | |
| I0506 01:39:27.162685 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 74.7365 > 30) by scale factor 0.40141 | |
| I0506 01:39:27.961073 15760 solver.cpp:229] Iteration 20500, loss = 9.83709 | |
| I0506 01:39:27.961163 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0217391 | |
| I0506 01:39:27.961182 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:39:27.961194 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0506 01:39:27.961207 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:39:27.961220 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:39:27.961231 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0506 01:39:27.961243 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0506 01:39:27.961256 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 01:39:27.961267 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:39:27.961283 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:39:27.961297 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0506 01:39:27.961308 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0506 01:39:27.961320 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0506 01:39:27.961333 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0506 01:39:27.961344 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:39:27.961356 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:39:27.961369 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:39:27.961380 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:39:27.961392 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:39:27.961403 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:39:27.961416 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:39:27.961427 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:39:27.961439 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:39:27.961450 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0506 01:39:27.961462 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.23913 | |
| I0506 01:39:27.961479 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.13336 (* 0.3 = 0.940008 loss) | |
| I0506 01:39:27.961493 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.912462 (* 0.3 = 0.273739 loss) | |
| I0506 01:39:27.961508 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.24785 (* 0.0272727 = 0.0885777 loss) | |
| I0506 01:39:27.961521 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.94835 (* 0.0272727 = 0.107682 loss) | |
| I0506 01:39:27.961535 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.44199 (* 0.0272727 = 0.0938725 loss) | |
| I0506 01:39:27.961549 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.49649 (* 0.0272727 = 0.0953589 loss) | |
| I0506 01:39:27.961562 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.16445 (* 0.0272727 = 0.0590306 loss) | |
| I0506 01:39:27.961575 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.55406 (* 0.0272727 = 0.0423836 loss) | |
| I0506 01:39:27.961590 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.688783 (* 0.0272727 = 0.018785 loss) | |
| I0506 01:39:27.961603 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.457752 (* 0.0272727 = 0.0124841 loss) | |
| I0506 01:39:27.961616 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.456465 (* 0.0272727 = 0.012449 loss) | |
| I0506 01:39:27.961630 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.5713 (* 0.0272727 = 0.0155809 loss) | |
| I0506 01:39:27.961643 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.504697 (* 0.0272727 = 0.0137645 loss) | |
| I0506 01:39:27.961658 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.48136 (* 0.0272727 = 0.013128 loss) | |
| I0506 01:39:27.961709 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.438532 (* 0.0272727 = 0.01196 loss) | |
| I0506 01:39:27.961724 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.111103 (* 0.0272727 = 0.00303008 loss) | |
| I0506 01:39:27.961738 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0646384 (* 0.0272727 = 0.00176287 loss) | |
| I0506 01:39:27.961752 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0303005 (* 0.0272727 = 0.000826378 loss) | |
| I0506 01:39:27.961766 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0164644 (* 0.0272727 = 0.000449028 loss) | |
| I0506 01:39:27.961781 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0055464 (* 0.0272727 = 0.000151265 loss) | |
| I0506 01:39:27.961793 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00827488 (* 0.0272727 = 0.000225678 loss) | |
| I0506 01:39:27.961807 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00531508 (* 0.0272727 = 0.000144957 loss) | |
| I0506 01:39:27.961822 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00651926 (* 0.0272727 = 0.000177798 loss) | |
| I0506 01:39:27.961834 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00330143 (* 0.0272727 = 9.00391e-05 loss) | |
| I0506 01:39:27.961846 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0217391 | |
| I0506 01:39:27.961858 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:39:27.961870 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:39:27.961881 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:39:27.961896 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:39:27.961908 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0506 01:39:27.961920 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0506 01:39:27.961932 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 01:39:27.961943 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 01:39:27.961956 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:39:27.961967 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0506 01:39:27.961979 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0506 01:39:27.961990 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0506 01:39:27.962002 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0506 01:39:27.962014 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:39:27.962025 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:39:27.962038 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:39:27.962049 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:39:27.962059 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:39:27.962071 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:39:27.962082 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:39:27.962093 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:39:27.962105 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:39:27.962116 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.738636 | |
| I0506 01:39:27.962128 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.26087 | |
| I0506 01:39:27.962142 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.10693 (* 0.3 = 0.932079 loss) | |
| I0506 01:39:27.962157 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.918064 (* 0.3 = 0.275419 loss) | |
| I0506 01:39:27.962170 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.99024 (* 0.0272727 = 0.081552 loss) | |
| I0506 01:39:27.962183 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.66492 (* 0.0272727 = 0.0999523 loss) | |
| I0506 01:39:27.962208 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.43211 (* 0.0272727 = 0.0936029 loss) | |
| I0506 01:39:27.962224 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.2673 (* 0.0272727 = 0.0891083 loss) | |
| I0506 01:39:27.962237 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.06973 (* 0.0272727 = 0.0564472 loss) | |
| I0506 01:39:27.962251 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.6978 (* 0.0272727 = 0.0463036 loss) | |
| I0506 01:39:27.962265 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.983129 (* 0.0272727 = 0.0268126 loss) | |
| I0506 01:39:27.962275 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.549641 (* 0.0272727 = 0.0149902 loss) | |
| I0506 01:39:27.962285 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.396262 (* 0.0272727 = 0.0108071 loss) | |
| I0506 01:39:27.962301 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.527998 (* 0.0272727 = 0.0144 loss) | |
| I0506 01:39:27.962314 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.56302 (* 0.0272727 = 0.0153551 loss) | |
| I0506 01:39:27.962328 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.721282 (* 0.0272727 = 0.0196713 loss) | |
| I0506 01:39:27.962347 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.555074 (* 0.0272727 = 0.0151384 loss) | |
| I0506 01:39:27.962362 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.034205 (* 0.0272727 = 0.000932865 loss) | |
| I0506 01:39:27.962375 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0542686 (* 0.0272727 = 0.00148005 loss) | |
| I0506 01:39:27.962389 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0232908 (* 0.0272727 = 0.000635203 loss) | |
| I0506 01:39:27.962404 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0147022 (* 0.0272727 = 0.000400968 loss) | |
| I0506 01:39:27.962416 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00889032 (* 0.0272727 = 0.000242463 loss) | |
| I0506 01:39:27.962430 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00555505 (* 0.0272727 = 0.000151501 loss) | |
| I0506 01:39:27.962445 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00594274 (* 0.0272727 = 0.000162075 loss) | |
| I0506 01:39:27.962457 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00628782 (* 0.0272727 = 0.000171486 loss) | |
| I0506 01:39:27.962471 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00785678 (* 0.0272727 = 0.000214276 loss) | |
| I0506 01:39:27.962483 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.108696 | |
| I0506 01:39:27.962496 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:39:27.962507 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0506 01:39:27.962519 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:39:27.962530 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0506 01:39:27.962543 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0506 01:39:27.962553 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0506 01:39:27.962565 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0506 01:39:27.962576 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 01:39:27.962589 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:39:27.962599 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0506 01:39:27.962610 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0506 01:39:27.962622 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0506 01:39:27.962633 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0506 01:39:27.962644 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:39:27.962656 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:39:27.962677 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:39:27.962690 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:39:27.962702 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:39:27.962713 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:39:27.962724 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:39:27.962735 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:39:27.962748 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:39:27.962759 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0506 01:39:27.962769 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.282609 | |
| I0506 01:39:27.962784 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.96292 (* 1 = 2.96292 loss) | |
| I0506 01:39:27.962797 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.852065 (* 1 = 0.852065 loss) | |
| I0506 01:39:27.962811 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.77656 (* 0.0909091 = 0.252414 loss) | |
| I0506 01:39:27.962826 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.17273 (* 0.0909091 = 0.28843 loss) | |
| I0506 01:39:27.962839 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.2019 (* 0.0909091 = 0.291082 loss) | |
| I0506 01:39:27.962852 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.02828 (* 0.0909091 = 0.275299 loss) | |
| I0506 01:39:27.962867 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 1.46116 (* 0.0909091 = 0.132832 loss) | |
| I0506 01:39:27.962879 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.33182 (* 0.0909091 = 0.121075 loss) | |
| I0506 01:39:27.962893 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.728917 (* 0.0909091 = 0.0662652 loss) | |
| I0506 01:39:27.962908 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.393579 (* 0.0909091 = 0.0357799 loss) | |
| I0506 01:39:27.962920 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.312316 (* 0.0909091 = 0.0283923 loss) | |
| I0506 01:39:27.962934 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.390745 (* 0.0909091 = 0.0355223 loss) | |
| I0506 01:39:27.962951 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.32169 (* 0.0909091 = 0.0292446 loss) | |
| I0506 01:39:27.962965 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.563257 (* 0.0909091 = 0.0512052 loss) | |
| I0506 01:39:27.962978 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.254756 (* 0.0909091 = 0.0231596 loss) | |
| I0506 01:39:27.962992 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0384251 (* 0.0909091 = 0.00349319 loss) | |
| I0506 01:39:27.963006 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.017981 (* 0.0909091 = 0.00163464 loss) | |
| I0506 01:39:27.963021 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00736314 (* 0.0909091 = 0.000669377 loss) | |
| I0506 01:39:27.963034 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0100271 (* 0.0909091 = 0.000911555 loss) | |
| I0506 01:39:27.963048 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00286412 (* 0.0909091 = 0.000260374 loss) | |
| I0506 01:39:27.963062 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00125714 (* 0.0909091 = 0.000114285 loss) | |
| I0506 01:39:27.963075 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00285668 (* 0.0909091 = 0.000259699 loss) | |
| I0506 01:39:27.963088 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00139374 (* 0.0909091 = 0.000126703 loss) | |
| I0506 01:39:27.963102 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00123529 (* 0.0909091 = 0.000112299 loss) | |
| I0506 01:39:27.963114 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:39:27.963125 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:39:27.963146 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000211799 | |
| I0506 01:39:27.963160 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000575102 | |
| I0506 01:39:27.963173 15760 sgd_solver.cpp:106] Iteration 20500, lr = 0.001 | |
| I0506 01:40:19.486096 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8864 > 30) by scale factor 0.971302 | |
| I0506 01:40:41.117801 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.1351 > 30) by scale factor 0.963543 | |
| I0506 01:40:44.964016 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.1935 > 30) by scale factor 0.728271 | |
| I0506 01:41:15.301951 15760 solver.cpp:229] Iteration 21000, loss = 9.86575 | |
| I0506 01:41:15.302078 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.075 | |
| I0506 01:41:15.302100 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:41:15.302114 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:41:15.302125 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:41:15.302137 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0506 01:41:15.302150 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:41:15.302175 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0506 01:41:15.302198 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0506 01:41:15.302212 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0506 01:41:15.302224 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:41:15.302237 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:41:15.302255 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:41:15.302266 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:41:15.302278 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:41:15.302289 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:41:15.302301 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:41:15.302314 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:41:15.302325 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:41:15.302336 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:41:15.302366 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:41:15.302377 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:41:15.302389 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:41:15.302400 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:41:15.302412 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.789773 | |
| I0506 01:41:15.302429 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.275 | |
| I0506 01:41:15.302445 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.95773 (* 0.3 = 0.887319 loss) | |
| I0506 01:41:15.302459 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.771874 (* 0.3 = 0.231562 loss) | |
| I0506 01:41:15.302474 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.23963 (* 0.0272727 = 0.0883537 loss) | |
| I0506 01:41:15.302487 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.05856 (* 0.0272727 = 0.0834154 loss) | |
| I0506 01:41:15.302501 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 2.98152 (* 0.0272727 = 0.0813141 loss) | |
| I0506 01:41:15.302515 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.126 (* 0.0272727 = 0.0579818 loss) | |
| I0506 01:41:15.302528 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.43531 (* 0.0272727 = 0.0664176 loss) | |
| I0506 01:41:15.302542 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.15865 (* 0.0272727 = 0.0588723 loss) | |
| I0506 01:41:15.302556 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 1.17099 (* 0.0272727 = 0.0319361 loss) | |
| I0506 01:41:15.302569 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0796253 (* 0.0272727 = 0.0021716 loss) | |
| I0506 01:41:15.302583 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0364238 (* 0.0272727 = 0.000993377 loss) | |
| I0506 01:41:15.302597 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.019524 (* 0.0272727 = 0.000532472 loss) | |
| I0506 01:41:15.302611 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0192795 (* 0.0272727 = 0.000525805 loss) | |
| I0506 01:41:15.302624 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0203192 (* 0.0272727 = 0.00055416 loss) | |
| I0506 01:41:15.302639 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0218101 (* 0.0272727 = 0.00059482 loss) | |
| I0506 01:41:15.302672 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0305868 (* 0.0272727 = 0.000834185 loss) | |
| I0506 01:41:15.302687 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0279883 (* 0.0272727 = 0.000763316 loss) | |
| I0506 01:41:15.302702 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.01604 (* 0.0272727 = 0.000437454 loss) | |
| I0506 01:41:15.302716 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0168131 (* 0.0272727 = 0.00045854 loss) | |
| I0506 01:41:15.302745 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0292577 (* 0.0272727 = 0.000797938 loss) | |
| I0506 01:41:15.302759 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0173683 (* 0.0272727 = 0.00047368 loss) | |
| I0506 01:41:15.302773 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0154449 (* 0.0272727 = 0.000421224 loss) | |
| I0506 01:41:15.302786 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00739256 (* 0.0272727 = 0.000201615 loss) | |
| I0506 01:41:15.302800 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0223862 (* 0.0272727 = 0.000610532 loss) | |
| I0506 01:41:15.302812 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.175 | |
| I0506 01:41:15.302825 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:41:15.302836 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0506 01:41:15.302848 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:41:15.302860 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.5 | |
| I0506 01:41:15.302871 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:41:15.302886 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0506 01:41:15.302898 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 01:41:15.302911 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0506 01:41:15.302922 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:41:15.302932 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:41:15.302943 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:41:15.302955 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:41:15.302966 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:41:15.302978 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:41:15.302989 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:41:15.302999 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:41:15.303010 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:41:15.303021 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:41:15.303032 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:41:15.303043 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:41:15.303055 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:41:15.303066 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:41:15.303077 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.801136 | |
| I0506 01:41:15.303092 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.325 | |
| I0506 01:41:15.303107 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.84484 (* 0.3 = 0.853451 loss) | |
| I0506 01:41:15.303120 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.76457 (* 0.3 = 0.229371 loss) | |
| I0506 01:41:15.303134 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.61084 (* 0.0272727 = 0.0712046 loss) | |
| I0506 01:41:15.303148 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.66361 (* 0.0272727 = 0.0726439 loss) | |
| I0506 01:41:15.303174 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 2.80314 (* 0.0272727 = 0.0764493 loss) | |
| I0506 01:41:15.303189 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 2.24684 (* 0.0272727 = 0.0612774 loss) | |
| I0506 01:41:15.303201 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.10996 (* 0.0272727 = 0.0575445 loss) | |
| I0506 01:41:15.303215 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 2.12213 (* 0.0272727 = 0.0578763 loss) | |
| I0506 01:41:15.303228 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.30065 (* 0.0272727 = 0.0354723 loss) | |
| I0506 01:41:15.303244 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0410543 (* 0.0272727 = 0.00111966 loss) | |
| I0506 01:41:15.303257 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0228115 (* 0.0272727 = 0.000622132 loss) | |
| I0506 01:41:15.303272 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0111743 (* 0.0272727 = 0.000304754 loss) | |
| I0506 01:41:15.303284 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0108795 (* 0.0272727 = 0.000296713 loss) | |
| I0506 01:41:15.303298 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00339825 (* 0.0272727 = 9.26797e-05 loss) | |
| I0506 01:41:15.303313 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00643621 (* 0.0272727 = 0.000175533 loss) | |
| I0506 01:41:15.303325 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00939647 (* 0.0272727 = 0.000256267 loss) | |
| I0506 01:41:15.303339 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0103929 (* 0.0272727 = 0.000283442 loss) | |
| I0506 01:41:15.303354 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00616458 (* 0.0272727 = 0.000168125 loss) | |
| I0506 01:41:15.303366 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00569456 (* 0.0272727 = 0.000155306 loss) | |
| I0506 01:41:15.303380 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00289212 (* 0.0272727 = 7.8876e-05 loss) | |
| I0506 01:41:15.303395 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00547476 (* 0.0272727 = 0.000149311 loss) | |
| I0506 01:41:15.303407 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00389556 (* 0.0272727 = 0.000106243 loss) | |
| I0506 01:41:15.303421 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0186206 (* 0.0272727 = 0.000507834 loss) | |
| I0506 01:41:15.303434 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0137633 (* 0.0272727 = 0.000375364 loss) | |
| I0506 01:41:15.303447 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.05 | |
| I0506 01:41:15.303458 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0506 01:41:15.303470 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:41:15.303481 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:41:15.303493 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.5 | |
| I0506 01:41:15.303503 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0506 01:41:15.303515 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0506 01:41:15.303526 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 01:41:15.303537 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0506 01:41:15.303550 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:41:15.303560 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:41:15.303571 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:41:15.303582 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:41:15.303593 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:41:15.303606 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:41:15.303617 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:41:15.303627 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:41:15.303648 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:41:15.303660 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:41:15.303671 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:41:15.303683 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:41:15.303694 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:41:15.303705 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:41:15.303716 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.784091 | |
| I0506 01:41:15.303727 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.3 | |
| I0506 01:41:15.303740 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.74939 (* 1 = 2.74939 loss) | |
| I0506 01:41:15.303755 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.68415 (* 1 = 0.68415 loss) | |
| I0506 01:41:15.303767 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.27775 (* 0.0909091 = 0.207068 loss) | |
| I0506 01:41:15.303781 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.47034 (* 0.0909091 = 0.224577 loss) | |
| I0506 01:41:15.303794 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.63664 (* 0.0909091 = 0.239694 loss) | |
| I0506 01:41:15.303807 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 1.8383 (* 0.0909091 = 0.167118 loss) | |
| I0506 01:41:15.303822 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.12734 (* 0.0909091 = 0.193394 loss) | |
| I0506 01:41:15.303834 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.72095 (* 0.0909091 = 0.15645 loss) | |
| I0506 01:41:15.303848 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 1.00963 (* 0.0909091 = 0.0917849 loss) | |
| I0506 01:41:15.303858 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0326375 (* 0.0909091 = 0.00296704 loss) | |
| I0506 01:41:15.303867 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00646006 (* 0.0909091 = 0.000587278 loss) | |
| I0506 01:41:15.303881 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.000933191 (* 0.0909091 = 8.48355e-05 loss) | |
| I0506 01:41:15.303895 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.000264759 (* 0.0909091 = 2.4069e-05 loss) | |
| I0506 01:41:15.303908 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.000203655 (* 0.0909091 = 1.85141e-05 loss) | |
| I0506 01:41:15.303922 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.000253192 (* 0.0909091 = 2.30175e-05 loss) | |
| I0506 01:41:15.303954 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000245214 (* 0.0909091 = 2.22922e-05 loss) | |
| I0506 01:41:15.303974 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000176224 (* 0.0909091 = 1.60204e-05 loss) | |
| I0506 01:41:15.303988 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000161661 (* 0.0909091 = 1.46965e-05 loss) | |
| I0506 01:41:15.304003 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000389183 (* 0.0909091 = 3.53803e-05 loss) | |
| I0506 01:41:15.304020 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000298864 (* 0.0909091 = 2.71695e-05 loss) | |
| I0506 01:41:15.304034 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00030509 (* 0.0909091 = 2.77355e-05 loss) | |
| I0506 01:41:15.304047 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000376152 (* 0.0909091 = 3.41956e-05 loss) | |
| I0506 01:41:15.304061 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000183993 (* 0.0909091 = 1.67266e-05 loss) | |
| I0506 01:41:15.304075 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000338962 (* 0.0909091 = 3.08148e-05 loss) | |
| I0506 01:41:15.304086 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:41:15.304097 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:41:15.304108 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000668356 | |
| I0506 01:41:15.304129 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00225692 | |
| I0506 01:41:15.304148 15760 sgd_solver.cpp:106] Iteration 21000, lr = 0.001 | |
| I0506 01:42:46.655160 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.2722 > 30) by scale factor 0.850529 | |
| I0506 01:42:55.474550 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.947 > 30) by scale factor 0.969398 | |
| I0506 01:43:03.180140 15760 solver.cpp:229] Iteration 21500, loss = 9.81008 | |
| I0506 01:43:03.180199 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.097561 | |
| I0506 01:43:03.180217 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0506 01:43:03.180229 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:43:03.180241 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0506 01:43:03.180260 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 01:43:03.180272 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:43:03.180284 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0506 01:43:03.180296 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 01:43:03.180308 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:43:03.180320 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:43:03.180335 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:43:03.180346 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:43:03.180357 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:43:03.180368 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:43:03.180380 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:43:03.180395 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:43:03.180418 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:43:03.180434 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:43:03.180464 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:43:03.180476 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:43:03.180487 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:43:03.180500 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:43:03.180510 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:43:03.180531 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.778409 | |
| I0506 01:43:03.180541 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.341463 | |
| I0506 01:43:03.180557 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.9943 (* 0.3 = 0.898291 loss) | |
| I0506 01:43:03.180572 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.82332 (* 0.3 = 0.246996 loss) | |
| I0506 01:43:03.180585 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.44943 (* 0.0272727 = 0.0940754 loss) | |
| I0506 01:43:03.180599 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.22774 (* 0.0272727 = 0.0880294 loss) | |
| I0506 01:43:03.180613 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.49019 (* 0.0272727 = 0.0951869 loss) | |
| I0506 01:43:03.180626 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.45204 (* 0.0272727 = 0.0941466 loss) | |
| I0506 01:43:03.180640 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.20408 (* 0.0272727 = 0.0601112 loss) | |
| I0506 01:43:03.180654 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.68988 (* 0.0272727 = 0.0460876 loss) | |
| I0506 01:43:03.180667 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.628578 (* 0.0272727 = 0.017143 loss) | |
| I0506 01:43:03.180681 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.5888 (* 0.0272727 = 0.0160582 loss) | |
| I0506 01:43:03.180696 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0437735 (* 0.0272727 = 0.00119382 loss) | |
| I0506 01:43:03.180711 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0343864 (* 0.0272727 = 0.000937811 loss) | |
| I0506 01:43:03.180734 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.02737 (* 0.0272727 = 0.000746454 loss) | |
| I0506 01:43:03.180797 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0169131 (* 0.0272727 = 0.000461266 loss) | |
| I0506 01:43:03.180814 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.011516 (* 0.0272727 = 0.000314072 loss) | |
| I0506 01:43:03.180826 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0108572 (* 0.0272727 = 0.000296105 loss) | |
| I0506 01:43:03.180850 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0122271 (* 0.0272727 = 0.000333466 loss) | |
| I0506 01:43:03.180865 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00651319 (* 0.0272727 = 0.000177633 loss) | |
| I0506 01:43:03.180877 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00539903 (* 0.0272727 = 0.000147246 loss) | |
| I0506 01:43:03.180892 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00361601 (* 0.0272727 = 9.86185e-05 loss) | |
| I0506 01:43:03.180907 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00346559 (* 0.0272727 = 9.45161e-05 loss) | |
| I0506 01:43:03.180920 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00517004 (* 0.0272727 = 0.000141001 loss) | |
| I0506 01:43:03.180933 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00183858 (* 0.0272727 = 5.01432e-05 loss) | |
| I0506 01:43:03.180948 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00275225 (* 0.0272727 = 7.50613e-05 loss) | |
| I0506 01:43:03.180959 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.146341 | |
| I0506 01:43:03.180971 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0506 01:43:03.180982 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:43:03.180994 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.375 | |
| I0506 01:43:03.181005 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0506 01:43:03.181016 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:43:03.181028 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0506 01:43:03.181040 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 01:43:03.181051 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 01:43:03.181062 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:43:03.181073 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:43:03.181085 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:43:03.181097 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:43:03.181107 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:43:03.181131 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:43:03.181146 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:43:03.181159 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:43:03.181169 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:43:03.181180 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:43:03.181191 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:43:03.181203 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:43:03.181215 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:43:03.181226 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:43:03.181236 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.795455 | |
| I0506 01:43:03.181248 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.341463 | |
| I0506 01:43:03.181265 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.98186 (* 0.3 = 0.894558 loss) | |
| I0506 01:43:03.181279 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.858501 (* 0.3 = 0.25755 loss) | |
| I0506 01:43:03.181293 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.74026 (* 0.0272727 = 0.102007 loss) | |
| I0506 01:43:03.181319 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 2.81876 (* 0.0272727 = 0.0768753 loss) | |
| I0506 01:43:03.181336 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 2.97142 (* 0.0272727 = 0.0810388 loss) | |
| I0506 01:43:03.181350 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.33204 (* 0.0272727 = 0.0908737 loss) | |
| I0506 01:43:03.181363 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.44306 (* 0.0272727 = 0.0666288 loss) | |
| I0506 01:43:03.181377 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.60262 (* 0.0272727 = 0.0437079 loss) | |
| I0506 01:43:03.181391 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.564649 (* 0.0272727 = 0.0153995 loss) | |
| I0506 01:43:03.181403 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.56775 (* 0.0272727 = 0.0154841 loss) | |
| I0506 01:43:03.181417 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.119758 (* 0.0272727 = 0.00326613 loss) | |
| I0506 01:43:03.181432 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.056474 (* 0.0272727 = 0.0015402 loss) | |
| I0506 01:43:03.181448 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.036609 (* 0.0272727 = 0.000998428 loss) | |
| I0506 01:43:03.181463 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0408065 (* 0.0272727 = 0.0011129 loss) | |
| I0506 01:43:03.181478 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0179319 (* 0.0272727 = 0.000489052 loss) | |
| I0506 01:43:03.181491 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0164141 (* 0.0272727 = 0.000447657 loss) | |
| I0506 01:43:03.181505 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0140822 (* 0.0272727 = 0.000384059 loss) | |
| I0506 01:43:03.181519 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00987764 (* 0.0272727 = 0.00026939 loss) | |
| I0506 01:43:03.181532 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00924986 (* 0.0272727 = 0.000252269 loss) | |
| I0506 01:43:03.181545 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00371083 (* 0.0272727 = 0.000101204 loss) | |
| I0506 01:43:03.181560 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00379429 (* 0.0272727 = 0.000103481 loss) | |
| I0506 01:43:03.181573 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00248551 (* 0.0272727 = 6.77866e-05 loss) | |
| I0506 01:43:03.181586 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00236915 (* 0.0272727 = 6.46131e-05 loss) | |
| I0506 01:43:03.181601 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00305078 (* 0.0272727 = 8.3203e-05 loss) | |
| I0506 01:43:03.181612 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0487805 | |
| I0506 01:43:03.181623 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:43:03.181635 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:43:03.181646 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:43:03.181658 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0506 01:43:03.181668 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0506 01:43:03.181680 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.75 | |
| I0506 01:43:03.181691 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0506 01:43:03.181702 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0506 01:43:03.181715 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0506 01:43:03.181726 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:43:03.181737 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:43:03.181753 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:43:03.181761 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:43:03.181774 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:43:03.181805 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:43:03.181826 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:43:03.181843 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:43:03.181859 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:43:03.181870 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:43:03.181881 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:43:03.181903 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:43:03.181915 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:43:03.181926 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0506 01:43:03.181938 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.268293 | |
| I0506 01:43:03.181951 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.92624 (* 1 = 2.92624 loss) | |
| I0506 01:43:03.181967 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.767335 (* 1 = 0.767335 loss) | |
| I0506 01:43:03.181993 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 3.36705 (* 0.0909091 = 0.306096 loss) | |
| I0506 01:43:03.182014 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 2.87342 (* 0.0909091 = 0.26122 loss) | |
| I0506 01:43:03.182029 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.80997 (* 0.0909091 = 0.255452 loss) | |
| I0506 01:43:03.182042 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.08379 (* 0.0909091 = 0.280344 loss) | |
| I0506 01:43:03.182055 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.00038 (* 0.0909091 = 0.181852 loss) | |
| I0506 01:43:03.182068 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 1.34176 (* 0.0909091 = 0.121979 loss) | |
| I0506 01:43:03.182087 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.601749 (* 0.0909091 = 0.0547045 loss) | |
| I0506 01:43:03.182101 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.531281 (* 0.0909091 = 0.0482983 loss) | |
| I0506 01:43:03.182114 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0717573 (* 0.0909091 = 0.00652339 loss) | |
| I0506 01:43:03.182128 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0278614 (* 0.0909091 = 0.00253286 loss) | |
| I0506 01:43:03.182142 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0138232 (* 0.0909091 = 0.00125665 loss) | |
| I0506 01:43:03.182154 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00661395 (* 0.0909091 = 0.000601268 loss) | |
| I0506 01:43:03.182168 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00534203 (* 0.0909091 = 0.000485639 loss) | |
| I0506 01:43:03.182181 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00587642 (* 0.0909091 = 0.00053422 loss) | |
| I0506 01:43:03.182194 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00382448 (* 0.0909091 = 0.00034768 loss) | |
| I0506 01:43:03.182207 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00331293 (* 0.0909091 = 0.000301175 loss) | |
| I0506 01:43:03.182226 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00337847 (* 0.0909091 = 0.000307134 loss) | |
| I0506 01:43:03.182240 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00195728 (* 0.0909091 = 0.000177934 loss) | |
| I0506 01:43:03.182252 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00165112 (* 0.0909091 = 0.000150102 loss) | |
| I0506 01:43:03.182266 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00120936 (* 0.0909091 = 0.000109942 loss) | |
| I0506 01:43:03.182279 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000927079 (* 0.0909091 = 8.42799e-05 loss) | |
| I0506 01:43:03.182301 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000644712 (* 0.0909091 = 5.86102e-05 loss) | |
| I0506 01:43:03.182312 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:43:03.182334 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:43:03.182348 15760 solver.cpp:245] Train net output #149: total_confidence = 0.000163064 | |
| I0506 01:43:03.182359 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000596317 | |
| I0506 01:43:03.182371 15760 sgd_solver.cpp:106] Iteration 21500, lr = 0.001 | |
| I0506 01:43:03.667918 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.8996 > 30) by scale factor 0.859609 | |
| I0506 01:43:36.488999 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 45.7652 > 30) by scale factor 0.655519 | |
| I0506 01:44:50.815320 15760 solver.cpp:229] Iteration 22000, loss = 9.86653 | |
| I0506 01:44:50.815454 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.04 | |
| I0506 01:44:50.815474 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:44:50.815488 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:44:50.815500 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0506 01:44:50.815512 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0506 01:44:50.815523 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 01:44:50.815536 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0506 01:44:50.815547 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 01:44:50.815559 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:44:50.815572 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0506 01:44:50.815583 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0506 01:44:50.815596 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:44:50.815608 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:44:50.815619 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:44:50.815631 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:44:50.815644 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:44:50.815656 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:44:50.815668 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:44:50.815680 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:44:50.815691 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:44:50.815703 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:44:50.815717 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:44:50.815738 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:44:50.815752 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.664773 | |
| I0506 01:44:50.815763 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.22 | |
| I0506 01:44:50.815779 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.18583 (* 0.3 = 0.955748 loss) | |
| I0506 01:44:50.815793 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.27078 (* 0.3 = 0.381234 loss) | |
| I0506 01:44:50.815809 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 3.47707 (* 0.0272727 = 0.0948292 loss) | |
| I0506 01:44:50.815822 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.97722 (* 0.0272727 = 0.10847 loss) | |
| I0506 01:44:50.815836 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.08723 (* 0.0272727 = 0.0841971 loss) | |
| I0506 01:44:50.815850 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.15651 (* 0.0272727 = 0.0860868 loss) | |
| I0506 01:44:50.815865 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.36026 (* 0.0272727 = 0.0916435 loss) | |
| I0506 01:44:50.815881 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 3.81015 (* 0.0272727 = 0.103913 loss) | |
| I0506 01:44:50.815896 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.800744 (* 0.0272727 = 0.0218385 loss) | |
| I0506 01:44:50.815910 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.961825 (* 0.0272727 = 0.0262316 loss) | |
| I0506 01:44:50.815924 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.795964 (* 0.0272727 = 0.0217081 loss) | |
| I0506 01:44:50.815938 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.959404 (* 0.0272727 = 0.0261656 loss) | |
| I0506 01:44:50.815951 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.546226 (* 0.0272727 = 0.0148971 loss) | |
| I0506 01:44:50.815965 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.436588 (* 0.0272727 = 0.0119069 loss) | |
| I0506 01:44:50.815979 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.381201 (* 0.0272727 = 0.0103964 loss) | |
| I0506 01:44:50.816011 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.248311 (* 0.0272727 = 0.00677211 loss) | |
| I0506 01:44:50.816026 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.255183 (* 0.0272727 = 0.00695952 loss) | |
| I0506 01:44:50.816040 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.110768 (* 0.0272727 = 0.00302094 loss) | |
| I0506 01:44:50.816054 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0431397 (* 0.0272727 = 0.00117654 loss) | |
| I0506 01:44:50.816069 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0354935 (* 0.0272727 = 0.000968004 loss) | |
| I0506 01:44:50.816083 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0388912 (* 0.0272727 = 0.00106067 loss) | |
| I0506 01:44:50.816097 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0132568 (* 0.0272727 = 0.00036155 loss) | |
| I0506 01:44:50.816112 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0144871 (* 0.0272727 = 0.000395103 loss) | |
| I0506 01:44:50.816125 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0133985 (* 0.0272727 = 0.000365414 loss) | |
| I0506 01:44:50.816138 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.08 | |
| I0506 01:44:50.816149 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:44:50.816161 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:44:50.816174 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:44:50.816185 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:44:50.816196 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0506 01:44:50.816207 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0506 01:44:50.816220 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0506 01:44:50.816231 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:44:50.816244 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0506 01:44:50.816256 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0506 01:44:50.816270 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:44:50.816290 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:44:50.816303 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:44:50.816314 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:44:50.816325 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:44:50.816337 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:44:50.816349 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:44:50.816359 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:44:50.816371 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:44:50.816382 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:44:50.816393 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:44:50.816404 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:44:50.816416 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0506 01:44:50.816428 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.24 | |
| I0506 01:44:50.816442 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.22309 (* 0.3 = 0.966926 loss) | |
| I0506 01:44:50.816457 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.14089 (* 0.3 = 0.342266 loss) | |
| I0506 01:44:50.816475 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 3.13866 (* 0.0272727 = 0.0855999 loss) | |
| I0506 01:44:50.816489 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.82128 (* 0.0272727 = 0.104217 loss) | |
| I0506 01:44:50.816504 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.26285 (* 0.0272727 = 0.0889869 loss) | |
| I0506 01:44:50.816529 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 2.99559 (* 0.0272727 = 0.081698 loss) | |
| I0506 01:44:50.816545 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.78013 (* 0.0272727 = 0.0758216 loss) | |
| I0506 01:44:50.816558 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.62302 (* 0.0272727 = 0.0988096 loss) | |
| I0506 01:44:50.816571 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 1.0807 (* 0.0272727 = 0.0294737 loss) | |
| I0506 01:44:50.816586 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.03257 (* 0.0272727 = 0.0281611 loss) | |
| I0506 01:44:50.816598 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.751611 (* 0.0272727 = 0.0204985 loss) | |
| I0506 01:44:50.816612 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.958176 (* 0.0272727 = 0.0261321 loss) | |
| I0506 01:44:50.816627 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.225473 (* 0.0272727 = 0.00614927 loss) | |
| I0506 01:44:50.816639 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.229921 (* 0.0272727 = 0.00627057 loss) | |
| I0506 01:44:50.816653 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.137808 (* 0.0272727 = 0.00375841 loss) | |
| I0506 01:44:50.816668 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0896624 (* 0.0272727 = 0.00244534 loss) | |
| I0506 01:44:50.816680 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.088586 (* 0.0272727 = 0.00241598 loss) | |
| I0506 01:44:50.816694 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0222397 (* 0.0272727 = 0.000606536 loss) | |
| I0506 01:44:50.816709 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0138269 (* 0.0272727 = 0.000377097 loss) | |
| I0506 01:44:50.816721 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00528553 (* 0.0272727 = 0.000144151 loss) | |
| I0506 01:44:50.816735 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00677551 (* 0.0272727 = 0.000184787 loss) | |
| I0506 01:44:50.816750 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00602799 (* 0.0272727 = 0.0001644 loss) | |
| I0506 01:44:50.816763 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00387946 (* 0.0272727 = 0.000105804 loss) | |
| I0506 01:44:50.816777 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00272527 (* 0.0272727 = 7.43255e-05 loss) | |
| I0506 01:44:50.816789 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.08 | |
| I0506 01:44:50.816802 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0506 01:44:50.816812 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:44:50.816823 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0506 01:44:50.816835 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:44:50.816846 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0506 01:44:50.816859 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0506 01:44:50.816869 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0506 01:44:50.816880 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:44:50.816892 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0506 01:44:50.816903 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0506 01:44:50.816915 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:44:50.816928 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:44:50.816941 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:44:50.816952 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:44:50.816963 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:44:50.816975 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:44:50.816997 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:44:50.817009 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:44:50.817020 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:44:50.817031 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:44:50.817044 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:44:50.817054 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:44:50.817065 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.698864 | |
| I0506 01:44:50.817077 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.32 | |
| I0506 01:44:50.817091 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.20712 (* 1 = 3.20712 loss) | |
| I0506 01:44:50.817104 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.16596 (* 1 = 1.16596 loss) | |
| I0506 01:44:50.817133 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.93878 (* 0.0909091 = 0.267162 loss) | |
| I0506 01:44:50.817152 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.60395 (* 0.0909091 = 0.327632 loss) | |
| I0506 01:44:50.817165 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 2.97407 (* 0.0909091 = 0.27037 loss) | |
| I0506 01:44:50.817178 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.12489 (* 0.0909091 = 0.284081 loss) | |
| I0506 01:44:50.817191 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 3.18527 (* 0.0909091 = 0.28957 loss) | |
| I0506 01:44:50.817205 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 3.50159 (* 0.0909091 = 0.318327 loss) | |
| I0506 01:44:50.817219 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 0.94783 (* 0.0909091 = 0.0861664 loss) | |
| I0506 01:44:50.817232 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 0.952377 (* 0.0909091 = 0.0865797 loss) | |
| I0506 01:44:50.817246 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.816229 (* 0.0909091 = 0.0742026 loss) | |
| I0506 01:44:50.817260 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.85622 (* 0.0909091 = 0.0778382 loss) | |
| I0506 01:44:50.817272 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.318728 (* 0.0909091 = 0.0289752 loss) | |
| I0506 01:44:50.817286 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.177642 (* 0.0909091 = 0.0161493 loss) | |
| I0506 01:44:50.817299 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.13148 (* 0.0909091 = 0.0119527 loss) | |
| I0506 01:44:50.817313 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0484561 (* 0.0909091 = 0.0044051 loss) | |
| I0506 01:44:50.817327 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0312042 (* 0.0909091 = 0.00283675 loss) | |
| I0506 01:44:50.817342 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0195665 (* 0.0909091 = 0.00177877 loss) | |
| I0506 01:44:50.817355 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0108038 (* 0.0909091 = 0.000982165 loss) | |
| I0506 01:44:50.817368 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00893583 (* 0.0909091 = 0.000812348 loss) | |
| I0506 01:44:50.817383 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00385931 (* 0.0909091 = 0.000350846 loss) | |
| I0506 01:44:50.817395 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00153983 (* 0.0909091 = 0.000139984 loss) | |
| I0506 01:44:50.817409 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00116841 (* 0.0909091 = 0.000106219 loss) | |
| I0506 01:44:50.817423 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000203797 (* 0.0909091 = 1.8527e-05 loss) | |
| I0506 01:44:50.817435 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:44:50.817446 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:44:50.817457 15760 solver.cpp:245] Train net output #149: total_confidence = 1.40205e-05 | |
| I0506 01:44:50.817481 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000125186 | |
| I0506 01:44:50.817495 15760 sgd_solver.cpp:106] Iteration 22000, lr = 0.001 | |
| I0506 01:44:56.303001 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 48.6849 > 30) by scale factor 0.616208 | |
| I0506 01:46:06.321776 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.7024 > 30) by scale factor 0.917364 | |
| I0506 01:46:25.498239 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 115.621 > 30) by scale factor 0.259468 | |
| I0506 01:46:38.783668 15760 solver.cpp:229] Iteration 22500, loss = 9.73912 | |
| I0506 01:46:38.783789 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0384615 | |
| I0506 01:46:38.783810 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0506 01:46:38.783823 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:46:38.783836 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0506 01:46:38.783849 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0506 01:46:38.783860 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0506 01:46:38.783872 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0506 01:46:38.783887 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0506 01:46:38.783900 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0506 01:46:38.783911 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0506 01:46:38.783923 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:46:38.783936 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:46:38.783946 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:46:38.783958 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:46:38.783969 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:46:38.783982 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:46:38.783993 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:46:38.784004 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:46:38.784016 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:46:38.784027 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:46:38.784039 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:46:38.784051 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:46:38.784062 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:46:38.784075 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.704545 | |
| I0506 01:46:38.784086 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.230769 | |
| I0506 01:46:38.784102 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.35521 (* 0.3 = 1.00656 loss) | |
| I0506 01:46:38.784116 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.12008 (* 0.3 = 0.336023 loss) | |
| I0506 01:46:38.784131 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.78806 (* 0.0272727 = 0.0760379 loss) | |
| I0506 01:46:38.784144 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 3.16936 (* 0.0272727 = 0.0864372 loss) | |
| I0506 01:46:38.784158 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.58785 (* 0.0272727 = 0.0978505 loss) | |
| I0506 01:46:38.784173 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 3.40414 (* 0.0272727 = 0.0928401 loss) | |
| I0506 01:46:38.784186 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 3.2479 (* 0.0272727 = 0.0885791 loss) | |
| I0506 01:46:38.784199 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 2.94696 (* 0.0272727 = 0.0803716 loss) | |
| I0506 01:46:38.784214 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 2.9754 (* 0.0272727 = 0.0811472 loss) | |
| I0506 01:46:38.784227 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 1.41444 (* 0.0272727 = 0.0385756 loss) | |
| I0506 01:46:38.784240 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.59087 (* 0.0272727 = 0.0161146 loss) | |
| I0506 01:46:38.784255 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.207092 (* 0.0272727 = 0.00564797 loss) | |
| I0506 01:46:38.784268 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.113819 (* 0.0272727 = 0.00310415 loss) | |
| I0506 01:46:38.784282 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.122696 (* 0.0272727 = 0.00334626 loss) | |
| I0506 01:46:38.784296 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.103288 (* 0.0272727 = 0.00281695 loss) | |
| I0506 01:46:38.784330 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.094634 (* 0.0272727 = 0.00258093 loss) | |
| I0506 01:46:38.784346 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0952354 (* 0.0272727 = 0.00259733 loss) | |
| I0506 01:46:38.784360 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0416378 (* 0.0272727 = 0.00113558 loss) | |
| I0506 01:46:38.784374 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0200568 (* 0.0272727 = 0.000547002 loss) | |
| I0506 01:46:38.784389 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0107033 (* 0.0272727 = 0.000291909 loss) | |
| I0506 01:46:38.784411 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00994478 (* 0.0272727 = 0.000271221 loss) | |
| I0506 01:46:38.784425 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0127035 (* 0.0272727 = 0.000346458 loss) | |
| I0506 01:46:38.784440 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00429126 (* 0.0272727 = 0.000117034 loss) | |
| I0506 01:46:38.784453 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00554698 (* 0.0272727 = 0.000151281 loss) | |
| I0506 01:46:38.784471 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0576923 | |
| I0506 01:46:38.784483 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:46:38.784495 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:46:38.784508 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0506 01:46:38.784519 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0506 01:46:38.784531 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0506 01:46:38.784543 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0506 01:46:38.784554 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0506 01:46:38.784565 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0506 01:46:38.784577 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0506 01:46:38.784590 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:46:38.784600 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:46:38.784612 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:46:38.784623 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:46:38.784634 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:46:38.784646 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:46:38.784657 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:46:38.784668 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:46:38.784679 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:46:38.784692 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:46:38.784703 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:46:38.784713 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:46:38.784725 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:46:38.784736 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.715909 | |
| I0506 01:46:38.784745 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.134615 | |
| I0506 01:46:38.784754 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.29507 (* 0.3 = 0.988522 loss) | |
| I0506 01:46:38.784770 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.05158 (* 0.3 = 0.315475 loss) | |
| I0506 01:46:38.784783 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.97221 (* 0.0272727 = 0.0810604 loss) | |
| I0506 01:46:38.784797 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.26853 (* 0.0272727 = 0.0891417 loss) | |
| I0506 01:46:38.784827 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.56099 (* 0.0272727 = 0.097118 loss) | |
| I0506 01:46:38.784843 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 3.16008 (* 0.0272727 = 0.0861839 loss) | |
| I0506 01:46:38.784857 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 3.47074 (* 0.0272727 = 0.0946565 loss) | |
| I0506 01:46:38.784870 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 3.09895 (* 0.0272727 = 0.0845169 loss) | |
| I0506 01:46:38.784884 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 2.76171 (* 0.0272727 = 0.0753193 loss) | |
| I0506 01:46:38.784898 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 1.30861 (* 0.0272727 = 0.0356895 loss) | |
| I0506 01:46:38.784911 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.601531 (* 0.0272727 = 0.0164054 loss) | |
| I0506 01:46:38.784927 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0931629 (* 0.0272727 = 0.00254081 loss) | |
| I0506 01:46:38.784942 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0773417 (* 0.0272727 = 0.00210932 loss) | |
| I0506 01:46:38.784956 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0767233 (* 0.0272727 = 0.00209245 loss) | |
| I0506 01:46:38.784970 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0904881 (* 0.0272727 = 0.00246786 loss) | |
| I0506 01:46:38.784983 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0469979 (* 0.0272727 = 0.00128176 loss) | |
| I0506 01:46:38.784997 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0313515 (* 0.0272727 = 0.000855041 loss) | |
| I0506 01:46:38.785012 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0451407 (* 0.0272727 = 0.00123111 loss) | |
| I0506 01:46:38.785024 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0362241 (* 0.0272727 = 0.000987929 loss) | |
| I0506 01:46:38.785038 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0245386 (* 0.0272727 = 0.000669234 loss) | |
| I0506 01:46:38.785053 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0315892 (* 0.0272727 = 0.000861524 loss) | |
| I0506 01:46:38.785065 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0083862 (* 0.0272727 = 0.000228715 loss) | |
| I0506 01:46:38.785079 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0106735 (* 0.0272727 = 0.000291096 loss) | |
| I0506 01:46:38.785092 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0148642 (* 0.0272727 = 0.000405388 loss) | |
| I0506 01:46:38.785104 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0769231 | |
| I0506 01:46:38.785116 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0506 01:46:38.785148 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:46:38.785161 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:46:38.785172 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0506 01:46:38.785183 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0506 01:46:38.785195 15760 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0506 01:46:38.785207 15760 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0506 01:46:38.785218 15760 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0506 01:46:38.785230 15760 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0506 01:46:38.785243 15760 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0506 01:46:38.785254 15760 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0506 01:46:38.785265 15760 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0506 01:46:38.785276 15760 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0506 01:46:38.785289 15760 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0506 01:46:38.785300 15760 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0506 01:46:38.785310 15760 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0506 01:46:38.785333 15760 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0506 01:46:38.785346 15760 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0506 01:46:38.785358 15760 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0506 01:46:38.785370 15760 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0506 01:46:38.785382 15760 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0506 01:46:38.785392 15760 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0506 01:46:38.785410 15760 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.710227 | |
| I0506 01:46:38.785423 15760 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.211538 | |
| I0506 01:46:38.785436 15760 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.18365 (* 1 = 3.18365 loss) | |
| I0506 01:46:38.785450 15760 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.06677 (* 1 = 1.06677 loss) | |
| I0506 01:46:38.785464 15760 solver.cpp:245] Train net output #125: loss3/loss01 = 2.70547 (* 0.0909091 = 0.245952 loss) | |
| I0506 01:46:38.785485 15760 solver.cpp:245] Train net output #126: loss3/loss02 = 3.2042 (* 0.0909091 = 0.291291 loss) | |
| I0506 01:46:38.785498 15760 solver.cpp:245] Train net output #127: loss3/loss03 = 3.39919 (* 0.0909091 = 0.309017 loss) | |
| I0506 01:46:38.785511 15760 solver.cpp:245] Train net output #128: loss3/loss04 = 3.34516 (* 0.0909091 = 0.304106 loss) | |
| I0506 01:46:38.785526 15760 solver.cpp:245] Train net output #129: loss3/loss05 = 2.96546 (* 0.0909091 = 0.269587 loss) | |
| I0506 01:46:38.785539 15760 solver.cpp:245] Train net output #130: loss3/loss06 = 2.90205 (* 0.0909091 = 0.263823 loss) | |
| I0506 01:46:38.785552 15760 solver.cpp:245] Train net output #131: loss3/loss07 = 2.47631 (* 0.0909091 = 0.225119 loss) | |
| I0506 01:46:38.785567 15760 solver.cpp:245] Train net output #132: loss3/loss08 = 1.04033 (* 0.0909091 = 0.0945753 loss) | |
| I0506 01:46:38.785580 15760 solver.cpp:245] Train net output #133: loss3/loss09 = 0.552452 (* 0.0909091 = 0.0502229 loss) | |
| I0506 01:46:38.785593 15760 solver.cpp:245] Train net output #134: loss3/loss10 = 0.102388 (* 0.0909091 = 0.009308 loss) | |
| I0506 01:46:38.785607 15760 solver.cpp:245] Train net output #135: loss3/loss11 = 0.081909 (* 0.0909091 = 0.00744627 loss) | |
| I0506 01:46:38.785621 15760 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0810213 (* 0.0909091 = 0.00736558 loss) | |
| I0506 01:46:38.785635 15760 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0786511 (* 0.0909091 = 0.0071501 loss) | |
| I0506 01:46:38.785648 15760 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0770173 (* 0.0909091 = 0.00700157 loss) | |
| I0506 01:46:38.785661 15760 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0631821 (* 0.0909091 = 0.00574383 loss) | |
| I0506 01:46:38.785676 15760 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0478215 (* 0.0909091 = 0.00434741 loss) | |
| I0506 01:46:38.785689 15760 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0373629 (* 0.0909091 = 0.00339663 loss) | |
| I0506 01:46:38.785702 15760 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0231246 (* 0.0909091 = 0.00210224 loss) | |
| I0506 01:46:38.785717 15760 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0314262 (* 0.0909091 = 0.00285693 loss) | |
| I0506 01:46:38.785729 15760 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0254509 (* 0.0909091 = 0.00231372 loss) | |
| I0506 01:46:38.785743 15760 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0187005 (* 0.0909091 = 0.00170004 loss) | |
| I0506 01:46:38.785756 15760 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0221344 (* 0.0909091 = 0.00201222 loss) | |
| I0506 01:46:38.785768 15760 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0506 01:46:38.785779 15760 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0506 01:46:38.785791 15760 solver.cpp:245] Train net output #149: total_confidence = 2.38985e-07 | |
| I0506 01:46:38.785812 15760 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.18112e-06 | |
| I0506 01:46:38.785826 15760 sgd_solver.cpp:106] Iteration 22500, lr = 0.001 | |
| I0506 01:46:55.903178 15760 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.6205 > 30) by scale factor 0.979736 | |
| I0506 01:48:27.095963 15760 solver.cpp:229] Iteration 23000, loss = 9.80171 | |
| I0506 01:48:27.096104 15760 solver.cpp:245] Train net output #0: loss1/accuracy = 0.102564 | |
| I0506 01:48:27.096127 15760 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0506 01:48:27.096139 15760 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0506 01:48:27.096151 15760 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0506 01:48:27.096163 15760 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0506 01:48:27.096175 15760 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0506 01:48:27.096187 15760 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0506 01:48:27.096199 15760 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0506 01:48:27.096210 15760 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0506 01:48:27.096223 15760 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0506 01:48:27.096235 15760 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0506 01:48:27.096246 15760 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0506 01:48:27.096258 15760 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0506 01:48:27.096269 15760 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0506 01:48:27.096282 15760 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0506 01:48:27.096293 15760 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0506 01:48:27.096305 15760 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0506 01:48:27.096318 15760 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0506 01:48:27.096328 15760 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0506 01:48:27.096340 15760 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0506 01:48:27.096351 15760 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0506 01:48:27.096364 15760 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0506 01:48:27.096374 15760 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0506 01:48:27.096386 15760 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.784091 | |
| I0506 01:48:27.096397 15760 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.205128 | |
| I0506 01:48:27.096415 15760 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.85816 (* 0.3 = 0.857448 loss) | |
| I0506 01:48:27.096428 15760 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.839456 (* 0.3 = 0.251837 loss) | |
| I0506 01:48:27.096443 15760 solver.cpp:245] Train net output #27: loss1/loss01 = 2.63258 (* 0.0272727 = 0.0717977 loss) | |
| I0506 01:48:27.096457 15760 solver.cpp:245] Train net output #28: loss1/loss02 = 2.96322 (* 0.0272727 = 0.080815 loss) | |
| I0506 01:48:27.096470 15760 solver.cpp:245] Train net output #29: loss1/loss03 = 3.20713 (* 0.0272727 = 0.0874672 loss) | |
| I0506 01:48:27.096484 15760 solver.cpp:245] Train net output #30: loss1/loss04 = 2.66459 (* 0.0272727 = 0.0726707 loss) | |
| I0506 01:48:27.096498 15760 solver.cpp:245] Train net output #31: loss1/loss05 = 2.1767 (* 0.0272727 = 0.0593646 loss) | |
| I0506 01:48:27.096511 15760 solver.cpp:245] Train net output #32: loss1/loss06 = 1.57929 (* 0.0272727 = 0.0430716 loss) | |
| I0506 01:48:27.096525 15760 solver.cpp:245] Train net output #33: loss1/loss07 = 0.751464 (* 0.0272727 = 0.0204945 loss) | |
| I0506 01:48:27.096539 15760 solver.cpp:245] Train net output #34: loss1/loss08 = 0.901534 (* 0.0272727 = 0.0245873 loss) | |
| I0506 01:48:27.096554 15760 solver.cpp:245] Train net output #35: loss1/loss09 = 0.109988 (* 0.0272727 = 0.00299968 loss) | |
| I0506 01:48:27.096567 15760 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0588343 (* 0.0272727 = 0.00160457 loss) | |
| I0506 01:48:27.096580 15760 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0595061 (* 0.0272727 = 0.00162289 loss) | |
| I0506 01:48:27.096603 15760 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0459467 (* 0.0272727 = 0.00125309 loss) | |
| I0506 01:48:27.096617 15760 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0361056 (* 0.0272727 = 0.000984698 loss) | |
| I0506 01:48:27.096650 15760 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0275801 (* 0.0272727 = 0.000752185 loss) | |
| I0506 01:48:27.096665 15760 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0237966 (* 0.0272727 = 0.000648998 loss) | |
| I0506 01:48:27.096679 15760 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0194347 (* 0.0272727 = 0.000530038 loss) | |
| I0506 01:48:27.096693 15760 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00678812 (* 0.0272727 = 0.00018513 loss) | |
| I0506 01:48:27.096707 15760 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00651316 (* 0.0272727 = 0.000177632 loss) | |
| I0506 01:48:27.096720 15760 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0102387 (* 0.0272727 = 0.000279237 loss) | |
| I0506 01:48:27.096735 15760 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0061118 (* 0.0272727 = 0.000166685 loss) | |
| I0506 01:48:27.096747 15760 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00419701 (* 0.0272727 = 0.000114464 loss) | |
| I0506 01:48:27.096761 15760 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00496545 (* 0.0272727 = 0.000135421 loss) | |
| I0506 01:48:27.096773 15760 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0512821 | |
| I0506 01:48:27.096786 15760 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0506 01:48:27.096806 15760 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0506 01:48:27.096817 15760 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0506 01:48:27.096828 15760 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0506 01:48:27.096839 15760 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0506 01:48:27.096848 15760 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.75 | |
| I0506 01:48:27.096855 15760 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0506 01:48:27.096879 15760 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0506 01:48:27.096891 15760 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0506 01:48:27.096904 15760 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0506 01:48:27.096915 15760 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0506 01:48:27.096925 15760 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0506 01:48:27.096936 15760 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0506 01:48:27.096947 15760 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0506 01:48:27.096959 15760 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0506 01:48:27.096971 15760 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0506 01:48:27.096982 15760 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0506 01:48:27.096992 15760 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0506 01:48:27.097003 15760 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0506 01:48:27.097015 15760 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0506 01:48:27.097026 15760 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0506 01:48:27.097038 15760 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0506 01:48:27.097048 15760 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0506 01:48:27.097060 15760 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.282051 | |
| I0506 01:48:27.097074 15760 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.82071 (* 0.3 = 0.846214 loss) | |
| I0506 01:48:27.097087 15760 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.798828 (* 0.3 = 0.239648 loss) | |
| I0506 01:48:27.097101 15760 solver.cpp:245] Train net output #76: loss2/loss01 = 2.54673 (* 0.0272727 = 0.0694562 loss) | |
| I0506 01:48:27.097128 15760 solver.cpp:245] Train net output #77: loss2/loss02 = 3.18762 (* 0.0272727 = 0.086935 loss) | |
| I0506 01:48:27.097158 15760 solver.cpp:245] Train net output #78: loss2/loss03 = 3.24409 (* 0.0272727 = 0.0884752 loss) | |
| I0506 01:48:27.097173 15760 solver.cpp:245] Train net output #79: loss2/loss04 = 2.88141 (* 0.0272727 = 0.0785839 loss) | |
| I0506 01:48:27.097187 15760 solver.cpp:245] Train net output #80: loss2/loss05 = 2.41184 (* 0.0272727 = 0.0657776 loss) | |
| I0506 01:48:27.097200 15760 solver.cpp:245] Train net output #81: loss2/loss06 = 1.47967 (* 0.0272727 = 0.0403545 loss) | |
| I0506 01:48:27.097214 15760 solver.cpp:245] Train net output #82: loss2/loss07 = 0.74732 (* 0.0272727 = 0.0203814 loss) | |
| I0506 01:48:27.097228 15760 solver.cpp:245] Train net output #83: loss2/loss08 = 0.790997 (* 0.0272727 = 0.0215726 loss) | |
| I0506 01:48:27.097241 15760 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0640428 (* 0.0272727 = 0.00174662 loss) | |
| I0506 01:48:27.097255 15760 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0313644 (* 0.0272727 = 0.000855392 loss) | |
| I0506 01:48:27.097270 15760 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0158407 (* 0.0272727 = 0.00043202 loss) | |
| I0506 01:48:27.097282 15760 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0247166 (* 0.0272727 = 0.000674088 loss) | |
| I0506 01:48:27.097296 15760 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0220872 (* 0.0272727 = 0.000602377 loss) | |
| I0506 01:48:27.097311 15760 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0165353 (* 0.0272727 = 0.000450962 loss) | |
| I0506 01:48:27.097323 15760 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0112308 (* 0.0272727 = 0.000306293 loss) | |
| I0506 01:48:27.097337 15760 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0096345 (* 0.0272727 = 0.000262759 loss) | |
| I0506 01:48:27.097352 15760 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00466891 (* 0.0272727 = 0.000127334 loss) | |
| I0506 01:48:27.097364 15760 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00339534 (* 0.0272727 = 9.26003e-05 loss) | |
| I0506 01:48:27.097378 15760 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00353045 (* 0.0272727 = 9.62849e-05 loss) | |
| I0506 01:48:27.097393 15760 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00182685 (* 0.0272727 = 4.98233e-05 loss) | |
| I0506 01:48:27.097406 15760 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00178632 (* 0.0272727 = 4.87177e-05 loss) | |
| I0506 01:48:27.097419 15760 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00256699 (* 0.0272727 = 7.00089e-05 loss) | |
| I0506 01:48:27.097431 15760 solver.cpp:245] Train net output #98: loss3/accuracy = 0.153846 | |
| I0506 01:48:27.097443 15760 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0506 01:48:27.097455 15760 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0506 01:48:27.097466 15760 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0506 01:48:27.097477 15760 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0506 01:48:27.097489 15760 solver.cpp:245] Train net output #103: loss3/accuracy05 = |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment