Last active
June 6, 2016 22:05
-
-
Save stas-sl/ce3346d38006f55793cd3d91cea6b9dc 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
| I0525 00:09:25.737826 5272 solver.cpp:280] Solving mixed_lstm | |
| I0525 00:09:25.737840 5272 solver.cpp:281] Learning Rate Policy: fixed | |
| I0525 00:09:26.273970 5272 solver.cpp:229] Iteration 0, loss = 27.9836 | |
| I0525 00:09:26.274058 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0525 00:09:26.274077 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 00:09:26.274091 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 00:09:26.274102 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 00:09:26.274114 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 00:09:26.274126 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 00:09:26.274138 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0 | |
| I0525 00:09:26.274149 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0 | |
| I0525 00:09:26.274161 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0 | |
| I0525 00:09:26.274173 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0 | |
| I0525 00:09:26.274184 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0 | |
| I0525 00:09:26.274195 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0 | |
| I0525 00:09:26.274207 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0 | |
| I0525 00:09:26.274219 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.125 | |
| I0525 00:09:26.274231 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.25 | |
| I0525 00:09:26.274243 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0 | |
| I0525 00:09:26.274255 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0 | |
| I0525 00:09:26.274266 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0 | |
| I0525 00:09:26.274277 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.125 | |
| I0525 00:09:26.274289 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0 | |
| I0525 00:09:26.274301 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 0 | |
| I0525 00:09:26.274312 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 0 | |
| I0525 00:09:26.274323 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 0 | |
| I0525 00:09:26.274334 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0 | |
| I0525 00:09:26.274346 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0416667 | |
| I0525 00:09:26.274363 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.42599 (* 0.3 = 1.3278 loss) | |
| I0525 00:09:26.274379 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 4.31819 (* 0.3 = 1.29546 loss) | |
| I0525 00:09:26.274392 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.69454 (* 0.0272727 = 0.128033 loss) | |
| I0525 00:09:26.274406 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.82997 (* 0.0272727 = 0.131726 loss) | |
| I0525 00:09:26.274420 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.90334 (* 0.0272727 = 0.133728 loss) | |
| I0525 00:09:26.274435 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 4.79926 (* 0.0272727 = 0.130889 loss) | |
| I0525 00:09:26.274448 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 4.85635 (* 0.0272727 = 0.132446 loss) | |
| I0525 00:09:26.274462 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 4.47805 (* 0.0272727 = 0.122129 loss) | |
| I0525 00:09:26.274477 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 4.62066 (* 0.0272727 = 0.126018 loss) | |
| I0525 00:09:26.274490 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 5.0867 (* 0.0272727 = 0.138728 loss) | |
| I0525 00:09:26.274504 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 4.28531 (* 0.0272727 = 0.116872 loss) | |
| I0525 00:09:26.274518 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 4.00745 (* 0.0272727 = 0.109294 loss) | |
| I0525 00:09:26.274533 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 5.00466 (* 0.0272727 = 0.136491 loss) | |
| I0525 00:09:26.274546 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 5.06219 (* 0.0272727 = 0.13806 loss) | |
| I0525 00:09:26.274559 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 3.84701 (* 0.0272727 = 0.104918 loss) | |
| I0525 00:09:26.274585 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 3.85242 (* 0.0272727 = 0.105066 loss) | |
| I0525 00:09:26.274600 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 4.94109 (* 0.0272727 = 0.134757 loss) | |
| I0525 00:09:26.274615 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 3.75635 (* 0.0272727 = 0.102446 loss) | |
| I0525 00:09:26.274627 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 3.97544 (* 0.0272727 = 0.108421 loss) | |
| I0525 00:09:26.274641 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 3.81998 (* 0.0272727 = 0.104181 loss) | |
| I0525 00:09:26.274655 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 4.77268 (* 0.0272727 = 0.130164 loss) | |
| I0525 00:09:26.274668 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 5.3693 (* 0.0272727 = 0.146435 loss) | |
| I0525 00:09:26.274682 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 4.20823 (* 0.0272727 = 0.11477 loss) | |
| I0525 00:09:26.274696 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 4.75394 (* 0.0272727 = 0.129653 loss) | |
| I0525 00:09:26.274708 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 00:09:26.274719 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 00:09:26.274731 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 00:09:26.274742 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 00:09:26.274757 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 00:09:26.274770 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0525 00:09:26.274780 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0525 00:09:26.274792 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0 | |
| I0525 00:09:26.274803 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0 | |
| I0525 00:09:26.274814 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0 | |
| I0525 00:09:26.274826 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0 | |
| I0525 00:09:26.274837 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0 | |
| I0525 00:09:26.274848 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0 | |
| I0525 00:09:26.274859 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0 | |
| I0525 00:09:26.274870 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0 | |
| I0525 00:09:26.274883 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0 | |
| I0525 00:09:26.274893 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.125 | |
| I0525 00:09:26.274904 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0 | |
| I0525 00:09:26.274916 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0 | |
| I0525 00:09:26.274927 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0 | |
| I0525 00:09:26.274938 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 0 | |
| I0525 00:09:26.274950 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 0 | |
| I0525 00:09:26.274960 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 0 | |
| I0525 00:09:26.274971 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0 | |
| I0525 00:09:26.274982 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0208333 | |
| I0525 00:09:26.274999 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.37339 (* 0.3 = 1.31202 loss) | |
| I0525 00:09:26.275013 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 4.56156 (* 0.3 = 1.36847 loss) | |
| I0525 00:09:26.275027 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.76261 (* 0.0272727 = 0.129889 loss) | |
| I0525 00:09:26.275041 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.69128 (* 0.0272727 = 0.127944 loss) | |
| I0525 00:09:26.275054 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.42595 (* 0.0272727 = 0.120708 loss) | |
| I0525 00:09:26.275079 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 5.29859 (* 0.0272727 = 0.144507 loss) | |
| I0525 00:09:26.275094 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 4.47793 (* 0.0272727 = 0.122125 loss) | |
| I0525 00:09:26.275109 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.95786 (* 0.0272727 = 0.107942 loss) | |
| I0525 00:09:26.275122 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 4.44144 (* 0.0272727 = 0.12113 loss) | |
| I0525 00:09:26.275136 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 4.40003 (* 0.0272727 = 0.120001 loss) | |
| I0525 00:09:26.275151 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 4.63225 (* 0.0272727 = 0.126334 loss) | |
| I0525 00:09:26.275163 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 4.60923 (* 0.0272727 = 0.125706 loss) | |
| I0525 00:09:26.275177 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 4.29059 (* 0.0272727 = 0.117016 loss) | |
| I0525 00:09:26.275192 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 4.41449 (* 0.0272727 = 0.120395 loss) | |
| I0525 00:09:26.275205 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 4.69665 (* 0.0272727 = 0.12809 loss) | |
| I0525 00:09:26.275219 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 3.86432 (* 0.0272727 = 0.105391 loss) | |
| I0525 00:09:26.275233 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 5.36989 (* 0.0272727 = 0.146452 loss) | |
| I0525 00:09:26.275248 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 3.81105 (* 0.0272727 = 0.103938 loss) | |
| I0525 00:09:26.275261 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 4.28254 (* 0.0272727 = 0.116797 loss) | |
| I0525 00:09:26.275274 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 5.43854 (* 0.0272727 = 0.148324 loss) | |
| I0525 00:09:26.275288 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 4.31123 (* 0.0272727 = 0.117579 loss) | |
| I0525 00:09:26.275302 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 4.3938 (* 0.0272727 = 0.119831 loss) | |
| I0525 00:09:26.275316 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 3.84154 (* 0.0272727 = 0.104769 loss) | |
| I0525 00:09:26.275331 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 4.09996 (* 0.0272727 = 0.111817 loss) | |
| I0525 00:09:26.275341 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0525 00:09:26.275353 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 00:09:26.275364 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:09:26.275377 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 00:09:26.275388 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 00:09:26.275399 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0525 00:09:26.275411 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0 | |
| I0525 00:09:26.275423 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0 | |
| I0525 00:09:26.275434 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0 | |
| I0525 00:09:26.275445 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0 | |
| I0525 00:09:26.275457 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0 | |
| I0525 00:09:26.275472 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0 | |
| I0525 00:09:26.275485 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0 | |
| I0525 00:09:26.275496 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0 | |
| I0525 00:09:26.275506 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0 | |
| I0525 00:09:26.275517 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0 | |
| I0525 00:09:26.275529 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0 | |
| I0525 00:09:26.275540 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0 | |
| I0525 00:09:26.275552 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0 | |
| I0525 00:09:26.275573 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0 | |
| I0525 00:09:26.275585 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 0 | |
| I0525 00:09:26.275598 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 0 | |
| I0525 00:09:26.275609 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 0 | |
| I0525 00:09:26.275620 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.0909091 | |
| I0525 00:09:26.275631 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0416667 | |
| I0525 00:09:26.275645 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.32039 (* 1 = 4.32039 loss) | |
| I0525 00:09:26.275658 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 4.11144 (* 1 = 4.11144 loss) | |
| I0525 00:09:26.275672 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 4.43148 (* 0.0909091 = 0.402862 loss) | |
| I0525 00:09:26.275687 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 4.33167 (* 0.0909091 = 0.393788 loss) | |
| I0525 00:09:26.275701 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.46903 (* 0.0909091 = 0.406276 loss) | |
| I0525 00:09:26.275714 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 4.53691 (* 0.0909091 = 0.412446 loss) | |
| I0525 00:09:26.275728 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 4.21558 (* 0.0909091 = 0.383234 loss) | |
| I0525 00:09:26.275741 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 4.92885 (* 0.0909091 = 0.448077 loss) | |
| I0525 00:09:26.275755 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 4.38094 (* 0.0909091 = 0.398267 loss) | |
| I0525 00:09:26.275769 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 4.81857 (* 0.0909091 = 0.438052 loss) | |
| I0525 00:09:26.275782 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 4.51415 (* 0.0909091 = 0.410377 loss) | |
| I0525 00:09:26.275799 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 4.28006 (* 0.0909091 = 0.389096 loss) | |
| I0525 00:09:26.275813 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 4.03323 (* 0.0909091 = 0.366658 loss) | |
| I0525 00:09:26.275827 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 4.22007 (* 0.0909091 = 0.383643 loss) | |
| I0525 00:09:26.275841 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 4.62234 (* 0.0909091 = 0.420213 loss) | |
| I0525 00:09:26.275854 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 4.29994 (* 0.0909091 = 0.390904 loss) | |
| I0525 00:09:26.275868 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 4.21356 (* 0.0909091 = 0.383051 loss) | |
| I0525 00:09:26.275882 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 4.16724 (* 0.0909091 = 0.37884 loss) | |
| I0525 00:09:26.275895 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 4.42567 (* 0.0909091 = 0.402334 loss) | |
| I0525 00:09:26.275909 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 4.65742 (* 0.0909091 = 0.423402 loss) | |
| I0525 00:09:26.275923 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 4.20004 (* 0.0909091 = 0.381822 loss) | |
| I0525 00:09:26.275938 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 4.21463 (* 0.0909091 = 0.383148 loss) | |
| I0525 00:09:26.275951 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 4.89986 (* 0.0909091 = 0.445442 loss) | |
| I0525 00:09:26.275965 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 4.33633 (* 0.0909091 = 0.394212 loss) | |
| I0525 00:09:26.275977 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:09:26.275988 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:09:26.276000 5272 solver.cpp:245] Train net output #149: total_confidence = 7.11833e-36 | |
| I0525 00:09:26.276011 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.88933e-31 | |
| I0525 00:09:26.276033 5272 sgd_solver.cpp:106] Iteration 0, lr = 0.001 | |
| I0525 00:15:51.341509 5272 solver.cpp:229] Iteration 500, loss = 14.5273 | |
| I0525 00:15:51.341969 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0525 00:15:51.341990 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 00:15:51.342003 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 00:15:51.342015 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 00:15:51.342028 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 00:15:51.342041 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 00:15:51.342051 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 00:15:51.342064 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 00:15:51.342077 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 00:15:51.342088 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 00:15:51.342099 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 00:15:51.342111 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 00:15:51.342123 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 00:15:51.342134 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 00:15:51.342145 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 00:15:51.342157 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 00:15:51.342170 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 00:15:51.342180 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 00:15:51.342191 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 00:15:51.342203 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:15:51.342216 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:15:51.342226 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:15:51.342238 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:15:51.342249 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0525 00:15:51.342262 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.155556 | |
| I0525 00:15:51.342278 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.16698 (* 0.3 = 1.25009 loss) | |
| I0525 00:15:51.342293 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.60984 (* 0.3 = 0.482952 loss) | |
| I0525 00:15:51.342308 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.93111 (* 0.0272727 = 0.107212 loss) | |
| I0525 00:15:51.342321 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.95271 (* 0.0272727 = 0.107801 loss) | |
| I0525 00:15:51.342335 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.36032 (* 0.0272727 = 0.118918 loss) | |
| I0525 00:15:51.342350 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.9613 (* 0.0272727 = 0.108035 loss) | |
| I0525 00:15:51.342363 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.75937 (* 0.0272727 = 0.102528 loss) | |
| I0525 00:15:51.342377 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.70828 (* 0.0272727 = 0.0738621 loss) | |
| I0525 00:15:51.342391 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 3.32993 (* 0.0272727 = 0.0908162 loss) | |
| I0525 00:15:51.342406 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.267894 (* 0.0272727 = 0.00730621 loss) | |
| I0525 00:15:51.342419 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.11113 (* 0.0272727 = 0.00303082 loss) | |
| I0525 00:15:51.342433 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.11995 (* 0.0272727 = 0.00327138 loss) | |
| I0525 00:15:51.342447 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.114153 (* 0.0272727 = 0.00311326 loss) | |
| I0525 00:15:51.342461 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0895019 (* 0.0272727 = 0.00244096 loss) | |
| I0525 00:15:51.342475 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0595278 (* 0.0272727 = 0.00162348 loss) | |
| I0525 00:15:51.342504 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0851157 (* 0.0272727 = 0.00232134 loss) | |
| I0525 00:15:51.342519 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0755571 (* 0.0272727 = 0.00206065 loss) | |
| I0525 00:15:51.342533 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0347603 (* 0.0272727 = 0.000948009 loss) | |
| I0525 00:15:51.342547 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0213723 (* 0.0272727 = 0.00058288 loss) | |
| I0525 00:15:51.342562 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0390867 (* 0.0272727 = 0.001066 loss) | |
| I0525 00:15:51.342576 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0495933 (* 0.0272727 = 0.00135255 loss) | |
| I0525 00:15:51.342591 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0303031 (* 0.0272727 = 0.000826448 loss) | |
| I0525 00:15:51.342604 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0191478 (* 0.0272727 = 0.000522212 loss) | |
| I0525 00:15:51.342618 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0457332 (* 0.0272727 = 0.00124727 loss) | |
| I0525 00:15:51.342631 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 00:15:51.342643 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 00:15:51.342654 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 00:15:51.342666 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 00:15:51.342677 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 00:15:51.342689 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 00:15:51.342701 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 00:15:51.342713 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 00:15:51.342721 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 00:15:51.342730 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 00:15:51.342741 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 00:15:51.342752 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 00:15:51.342764 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 00:15:51.342775 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 00:15:51.342787 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 00:15:51.342797 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 00:15:51.342809 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 00:15:51.342820 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 00:15:51.342831 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 00:15:51.342842 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:15:51.342854 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:15:51.342864 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:15:51.342878 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:15:51.342890 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0525 00:15:51.342902 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0444444 | |
| I0525 00:15:51.342916 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.27844 (* 0.3 = 1.28353 loss) | |
| I0525 00:15:51.342929 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.67829 (* 0.3 = 0.503488 loss) | |
| I0525 00:15:51.342943 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.85101 (* 0.0272727 = 0.105028 loss) | |
| I0525 00:15:51.342957 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.11976 (* 0.0272727 = 0.112357 loss) | |
| I0525 00:15:51.342970 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.15074 (* 0.0272727 = 0.113202 loss) | |
| I0525 00:15:51.343000 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 4.0261 (* 0.0272727 = 0.109803 loss) | |
| I0525 00:15:51.343016 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 4.13462 (* 0.0272727 = 0.112762 loss) | |
| I0525 00:15:51.343029 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.86264 (* 0.0272727 = 0.0780721 loss) | |
| I0525 00:15:51.343044 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.97172 (* 0.0272727 = 0.081047 loss) | |
| I0525 00:15:51.343057 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.361772 (* 0.0272727 = 0.00986651 loss) | |
| I0525 00:15:51.343071 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.240554 (* 0.0272727 = 0.00656058 loss) | |
| I0525 00:15:51.343086 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.362438 (* 0.0272727 = 0.00988467 loss) | |
| I0525 00:15:51.343099 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.148349 (* 0.0272727 = 0.00404588 loss) | |
| I0525 00:15:51.343113 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0827182 (* 0.0272727 = 0.00225595 loss) | |
| I0525 00:15:51.343127 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.12439 (* 0.0272727 = 0.00339247 loss) | |
| I0525 00:15:51.343142 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.116527 (* 0.0272727 = 0.003178 loss) | |
| I0525 00:15:51.343154 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.105147 (* 0.0272727 = 0.00286766 loss) | |
| I0525 00:15:51.343168 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.059241 (* 0.0272727 = 0.00161566 loss) | |
| I0525 00:15:51.343183 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0836293 (* 0.0272727 = 0.0022808 loss) | |
| I0525 00:15:51.343196 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0844754 (* 0.0272727 = 0.00230387 loss) | |
| I0525 00:15:51.343210 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.03367 (* 0.0272727 = 0.000918272 loss) | |
| I0525 00:15:51.343225 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.052629 (* 0.0272727 = 0.00143534 loss) | |
| I0525 00:15:51.343237 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0451705 (* 0.0272727 = 0.00123192 loss) | |
| I0525 00:15:51.343252 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0656845 (* 0.0272727 = 0.0017914 loss) | |
| I0525 00:15:51.343264 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0525 00:15:51.343276 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 00:15:51.343287 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:15:51.343298 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 00:15:51.343309 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 00:15:51.343322 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 00:15:51.343333 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 00:15:51.343344 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 00:15:51.343355 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 00:15:51.343366 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 00:15:51.343379 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 00:15:51.343389 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 00:15:51.343400 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 00:15:51.343411 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 00:15:51.343422 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 00:15:51.343433 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 00:15:51.343446 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 00:15:51.343456 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 00:15:51.343487 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 00:15:51.343510 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:15:51.343524 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:15:51.343536 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:15:51.343547 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:15:51.343559 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.738636 | |
| I0525 00:15:51.343570 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.133333 | |
| I0525 00:15:51.343585 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.75694 (* 1 = 3.75694 loss) | |
| I0525 00:15:51.343598 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.20916 (* 1 = 1.20916 loss) | |
| I0525 00:15:51.343612 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.76891 (* 0.0909091 = 0.342629 loss) | |
| I0525 00:15:51.343626 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.48806 (* 0.0909091 = 0.317096 loss) | |
| I0525 00:15:51.343639 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.05613 (* 0.0909091 = 0.368739 loss) | |
| I0525 00:15:51.343653 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.34506 (* 0.0909091 = 0.304096 loss) | |
| I0525 00:15:51.343667 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.1946 (* 0.0909091 = 0.290418 loss) | |
| I0525 00:15:51.343682 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.23188 (* 0.0909091 = 0.202898 loss) | |
| I0525 00:15:51.343695 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 3.04774 (* 0.0909091 = 0.277067 loss) | |
| I0525 00:15:51.343709 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.201753 (* 0.0909091 = 0.0183412 loss) | |
| I0525 00:15:51.343722 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.122778 (* 0.0909091 = 0.0111616 loss) | |
| I0525 00:15:51.343736 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.08909 (* 0.0909091 = 0.00809909 loss) | |
| I0525 00:15:51.343750 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0940597 (* 0.0909091 = 0.00855088 loss) | |
| I0525 00:15:51.343765 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0746214 (* 0.0909091 = 0.00678376 loss) | |
| I0525 00:15:51.343778 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0745911 (* 0.0909091 = 0.00678101 loss) | |
| I0525 00:15:51.343792 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0355917 (* 0.0909091 = 0.00323561 loss) | |
| I0525 00:15:51.343806 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0508557 (* 0.0909091 = 0.00462325 loss) | |
| I0525 00:15:51.343819 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0291794 (* 0.0909091 = 0.00265267 loss) | |
| I0525 00:15:51.343833 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.012849 (* 0.0909091 = 0.00116809 loss) | |
| I0525 00:15:51.343847 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0118228 (* 0.0909091 = 0.0010748 loss) | |
| I0525 00:15:51.343861 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00941348 (* 0.0909091 = 0.00085577 loss) | |
| I0525 00:15:51.343875 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0074061 (* 0.0909091 = 0.000673282 loss) | |
| I0525 00:15:51.343889 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00801268 (* 0.0909091 = 0.000728426 loss) | |
| I0525 00:15:51.343904 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00776598 (* 0.0909091 = 0.000705999 loss) | |
| I0525 00:15:51.343915 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:15:51.343930 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:15:51.343941 5272 solver.cpp:245] Train net output #149: total_confidence = 1.41157e-08 | |
| I0525 00:15:51.343953 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.65183e-06 | |
| I0525 00:15:51.343976 5272 sgd_solver.cpp:106] Iteration 500, lr = 0.001 | |
| I0525 00:22:16.113615 5272 solver.cpp:229] Iteration 1000, loss = 13.4113 | |
| I0525 00:22:16.113762 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0525 00:22:16.113782 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 00:22:16.113796 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0525 00:22:16.113808 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 00:22:16.113821 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 00:22:16.113832 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.625 | |
| I0525 00:22:16.113845 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 00:22:16.113857 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 00:22:16.113868 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 00:22:16.113883 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 00:22:16.113895 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 00:22:16.113906 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 00:22:16.113919 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 00:22:16.113930 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 00:22:16.113941 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 00:22:16.113953 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 00:22:16.113965 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 00:22:16.113976 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 00:22:16.113988 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 00:22:16.113999 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:22:16.114012 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:22:16.114022 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:22:16.114034 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:22:16.114045 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.784091 | |
| I0525 00:22:16.114058 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0526316 | |
| I0525 00:22:16.114074 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.98246 (* 0.3 = 1.19474 loss) | |
| I0525 00:22:16.114089 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.40073 (* 0.3 = 0.420219 loss) | |
| I0525 00:22:16.114104 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.16377 (* 0.0272727 = 0.113557 loss) | |
| I0525 00:22:16.114117 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.42226 (* 0.0272727 = 0.0933343 loss) | |
| I0525 00:22:16.114131 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.70956 (* 0.0272727 = 0.10117 loss) | |
| I0525 00:22:16.114145 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.75778 (* 0.0272727 = 0.102485 loss) | |
| I0525 00:22:16.114159 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.78501 (* 0.0272727 = 0.0759548 loss) | |
| I0525 00:22:16.114173 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.49968 (* 0.0272727 = 0.0681731 loss) | |
| I0525 00:22:16.114187 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.87425 (* 0.0272727 = 0.0511159 loss) | |
| I0525 00:22:16.114200 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.371593 (* 0.0272727 = 0.0101343 loss) | |
| I0525 00:22:16.114214 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.147338 (* 0.0272727 = 0.0040183 loss) | |
| I0525 00:22:16.114228 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.260484 (* 0.0272727 = 0.0071041 loss) | |
| I0525 00:22:16.114243 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.101823 (* 0.0272727 = 0.002777 loss) | |
| I0525 00:22:16.114258 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.101002 (* 0.0272727 = 0.00275459 loss) | |
| I0525 00:22:16.114271 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.081603 (* 0.0272727 = 0.00222554 loss) | |
| I0525 00:22:16.114305 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.05648 (* 0.0272727 = 0.00154036 loss) | |
| I0525 00:22:16.114320 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0845832 (* 0.0272727 = 0.00230682 loss) | |
| I0525 00:22:16.114333 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0476369 (* 0.0272727 = 0.00129919 loss) | |
| I0525 00:22:16.114347 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0586945 (* 0.0272727 = 0.00160076 loss) | |
| I0525 00:22:16.114362 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.021597 (* 0.0272727 = 0.000589008 loss) | |
| I0525 00:22:16.114375 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0494823 (* 0.0272727 = 0.00134952 loss) | |
| I0525 00:22:16.114389 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0448972 (* 0.0272727 = 0.00122447 loss) | |
| I0525 00:22:16.114403 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0443787 (* 0.0272727 = 0.00121033 loss) | |
| I0525 00:22:16.114416 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0291346 (* 0.0272727 = 0.00079458 loss) | |
| I0525 00:22:16.114429 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 00:22:16.114441 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 00:22:16.114452 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 00:22:16.114464 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 00:22:16.114475 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 00:22:16.114486 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.625 | |
| I0525 00:22:16.114498 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 00:22:16.114509 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 00:22:16.114521 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 00:22:16.114532 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 00:22:16.114543 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 00:22:16.114555 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 00:22:16.114567 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 00:22:16.114578 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 00:22:16.114589 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 00:22:16.114598 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 00:22:16.114604 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 00:22:16.114617 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 00:22:16.114629 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 00:22:16.114639 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:22:16.114650 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:22:16.114662 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:22:16.114673 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:22:16.114684 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.784091 | |
| I0525 00:22:16.114696 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0263158 | |
| I0525 00:22:16.114709 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.00513 (* 0.3 = 1.20154 loss) | |
| I0525 00:22:16.114722 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.33065 (* 0.3 = 0.399195 loss) | |
| I0525 00:22:16.114737 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.86058 (* 0.0272727 = 0.105288 loss) | |
| I0525 00:22:16.114750 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.13858 (* 0.0272727 = 0.11287 loss) | |
| I0525 00:22:16.114763 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.37234 (* 0.0272727 = 0.119246 loss) | |
| I0525 00:22:16.114792 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.73512 (* 0.0272727 = 0.101867 loss) | |
| I0525 00:22:16.114807 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.40307 (* 0.0272727 = 0.0655383 loss) | |
| I0525 00:22:16.114821 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 1.99799 (* 0.0272727 = 0.0544907 loss) | |
| I0525 00:22:16.114835 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.57227 (* 0.0272727 = 0.04288 loss) | |
| I0525 00:22:16.114850 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.199847 (* 0.0272727 = 0.00545038 loss) | |
| I0525 00:22:16.114863 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0967752 (* 0.0272727 = 0.00263932 loss) | |
| I0525 00:22:16.114877 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.104619 (* 0.0272727 = 0.00285325 loss) | |
| I0525 00:22:16.114891 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0818492 (* 0.0272727 = 0.00223225 loss) | |
| I0525 00:22:16.114904 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.10203 (* 0.0272727 = 0.00278264 loss) | |
| I0525 00:22:16.114918 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.172598 (* 0.0272727 = 0.00470721 loss) | |
| I0525 00:22:16.114935 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0475379 (* 0.0272727 = 0.00129649 loss) | |
| I0525 00:22:16.114948 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0649334 (* 0.0272727 = 0.00177091 loss) | |
| I0525 00:22:16.114962 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0437856 (* 0.0272727 = 0.00119415 loss) | |
| I0525 00:22:16.114976 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.027479 (* 0.0272727 = 0.000749428 loss) | |
| I0525 00:22:16.114990 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.035191 (* 0.0272727 = 0.000959754 loss) | |
| I0525 00:22:16.115003 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.034326 (* 0.0272727 = 0.000936163 loss) | |
| I0525 00:22:16.115017 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0263038 (* 0.0272727 = 0.000717375 loss) | |
| I0525 00:22:16.115031 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0461519 (* 0.0272727 = 0.00125869 loss) | |
| I0525 00:22:16.115044 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0367581 (* 0.0272727 = 0.00100249 loss) | |
| I0525 00:22:16.115056 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0789474 | |
| I0525 00:22:16.115068 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 00:22:16.115079 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 00:22:16.115092 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 00:22:16.115103 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 00:22:16.115114 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0525 00:22:16.115125 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 00:22:16.115137 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 00:22:16.115149 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 00:22:16.115160 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 00:22:16.115171 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 00:22:16.115182 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 00:22:16.115195 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 00:22:16.115206 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 00:22:16.115216 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 00:22:16.115228 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 00:22:16.115239 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 00:22:16.115250 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 00:22:16.115270 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 00:22:16.115283 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:22:16.115295 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:22:16.115306 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:22:16.115317 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:22:16.115329 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.789773 | |
| I0525 00:22:16.115340 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.105263 | |
| I0525 00:22:16.115355 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.60232 (* 1 = 3.60232 loss) | |
| I0525 00:22:16.115367 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.08289 (* 1 = 1.08289 loss) | |
| I0525 00:22:16.115383 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.58339 (* 0.0909091 = 0.325763 loss) | |
| I0525 00:22:16.115411 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.46167 (* 0.0909091 = 0.314697 loss) | |
| I0525 00:22:16.115434 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.69659 (* 0.0909091 = 0.336054 loss) | |
| I0525 00:22:16.115449 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.08156 (* 0.0909091 = 0.280142 loss) | |
| I0525 00:22:16.115463 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.60997 (* 0.0909091 = 0.23727 loss) | |
| I0525 00:22:16.115476 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.99255 (* 0.0909091 = 0.181141 loss) | |
| I0525 00:22:16.115490 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.61895 (* 0.0909091 = 0.147177 loss) | |
| I0525 00:22:16.115504 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.243698 (* 0.0909091 = 0.0221543 loss) | |
| I0525 00:22:16.115517 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.114596 (* 0.0909091 = 0.0104178 loss) | |
| I0525 00:22:16.115531 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.14765 (* 0.0909091 = 0.0134228 loss) | |
| I0525 00:22:16.115545 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.122678 (* 0.0909091 = 0.0111525 loss) | |
| I0525 00:22:16.115559 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.132164 (* 0.0909091 = 0.0120149 loss) | |
| I0525 00:22:16.115573 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0639484 (* 0.0909091 = 0.00581349 loss) | |
| I0525 00:22:16.115586 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0459391 (* 0.0909091 = 0.00417628 loss) | |
| I0525 00:22:16.115602 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0406054 (* 0.0909091 = 0.0036914 loss) | |
| I0525 00:22:16.115615 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0210758 (* 0.0909091 = 0.00191599 loss) | |
| I0525 00:22:16.115628 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0137746 (* 0.0909091 = 0.00125223 loss) | |
| I0525 00:22:16.115643 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0102468 (* 0.0909091 = 0.000931525 loss) | |
| I0525 00:22:16.115656 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00757501 (* 0.0909091 = 0.000688637 loss) | |
| I0525 00:22:16.115670 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00679802 (* 0.0909091 = 0.000618002 loss) | |
| I0525 00:22:16.115684 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00615695 (* 0.0909091 = 0.000559722 loss) | |
| I0525 00:22:16.115697 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00554393 (* 0.0909091 = 0.000503994 loss) | |
| I0525 00:22:16.115710 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:22:16.115720 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:22:16.115731 5272 solver.cpp:245] Train net output #149: total_confidence = 4.67072e-09 | |
| I0525 00:22:16.115743 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.42419e-06 | |
| I0525 00:22:16.115767 5272 sgd_solver.cpp:106] Iteration 1000, lr = 0.001 | |
| I0525 00:28:40.750103 5272 solver.cpp:229] Iteration 1500, loss = 13.1165 | |
| I0525 00:28:40.750239 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0525 00:28:40.750264 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 00:28:40.750279 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 00:28:40.750291 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 00:28:40.750303 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 00:28:40.750315 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 00:28:40.750327 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 00:28:40.750339 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0525 00:28:40.750351 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0525 00:28:40.750368 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0525 00:28:40.750380 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0525 00:28:40.750392 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.75 | |
| I0525 00:28:40.750404 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 00:28:40.750416 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 00:28:40.750429 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 00:28:40.750440 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 00:28:40.750452 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 00:28:40.750463 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 00:28:40.750483 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 00:28:40.750496 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:28:40.750507 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:28:40.750519 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:28:40.750530 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:28:40.750541 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.659091 | |
| I0525 00:28:40.750553 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0666667 | |
| I0525 00:28:40.750569 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.05514 (* 0.3 = 1.21654 loss) | |
| I0525 00:28:40.750584 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.68482 (* 0.3 = 0.505445 loss) | |
| I0525 00:28:40.750598 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.79241 (* 0.0272727 = 0.103429 loss) | |
| I0525 00:28:40.750612 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.68147 (* 0.0272727 = 0.100404 loss) | |
| I0525 00:28:40.750627 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.17152 (* 0.0272727 = 0.113769 loss) | |
| I0525 00:28:40.750639 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.5854 (* 0.0272727 = 0.0977836 loss) | |
| I0525 00:28:40.750653 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.68295 (* 0.0272727 = 0.100444 loss) | |
| I0525 00:28:40.750668 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.15246 (* 0.0272727 = 0.0859761 loss) | |
| I0525 00:28:40.750680 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.92764 (* 0.0272727 = 0.0798448 loss) | |
| I0525 00:28:40.750694 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 2.29493 (* 0.0272727 = 0.062589 loss) | |
| I0525 00:28:40.750708 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.80476 (* 0.0272727 = 0.0492208 loss) | |
| I0525 00:28:40.750722 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 1.66005 (* 0.0272727 = 0.0452742 loss) | |
| I0525 00:28:40.750736 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 1.78289 (* 0.0272727 = 0.0486242 loss) | |
| I0525 00:28:40.750757 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.888369 (* 0.0272727 = 0.0242282 loss) | |
| I0525 00:28:40.750772 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.828803 (* 0.0272727 = 0.0226037 loss) | |
| I0525 00:28:40.750805 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.820904 (* 0.0272727 = 0.0223883 loss) | |
| I0525 00:28:40.750820 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 1.08018 (* 0.0272727 = 0.0294596 loss) | |
| I0525 00:28:40.750834 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0363395 (* 0.0272727 = 0.000991077 loss) | |
| I0525 00:28:40.750849 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0233024 (* 0.0272727 = 0.00063552 loss) | |
| I0525 00:28:40.750862 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0220617 (* 0.0272727 = 0.000601682 loss) | |
| I0525 00:28:40.750876 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.020704 (* 0.0272727 = 0.000564653 loss) | |
| I0525 00:28:40.750890 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0165594 (* 0.0272727 = 0.000451619 loss) | |
| I0525 00:28:40.750905 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0203279 (* 0.0272727 = 0.000554397 loss) | |
| I0525 00:28:40.750917 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0282538 (* 0.0272727 = 0.000770559 loss) | |
| I0525 00:28:40.750929 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 00:28:40.750941 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 00:28:40.750952 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 00:28:40.750963 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 00:28:40.750974 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 00:28:40.750987 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 00:28:40.750998 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 00:28:40.751009 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0525 00:28:40.751020 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0525 00:28:40.751031 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0525 00:28:40.751044 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0525 00:28:40.751055 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.75 | |
| I0525 00:28:40.751065 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 00:28:40.751077 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 00:28:40.751088 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 00:28:40.751101 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 00:28:40.751111 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 00:28:40.751122 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 00:28:40.751133 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 00:28:40.751144 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:28:40.751155 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:28:40.751166 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:28:40.751178 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:28:40.751188 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.659091 | |
| I0525 00:28:40.751200 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0833333 | |
| I0525 00:28:40.751214 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.1402 (* 0.3 = 1.24206 loss) | |
| I0525 00:28:40.751226 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.74291 (* 0.3 = 0.522874 loss) | |
| I0525 00:28:40.751240 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.22948 (* 0.0272727 = 0.115349 loss) | |
| I0525 00:28:40.751255 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.16168 (* 0.0272727 = 0.1135 loss) | |
| I0525 00:28:40.751278 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.19356 (* 0.0272727 = 0.11437 loss) | |
| I0525 00:28:40.751292 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.67371 (* 0.0272727 = 0.100192 loss) | |
| I0525 00:28:40.751307 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.78463 (* 0.0272727 = 0.103217 loss) | |
| I0525 00:28:40.751320 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.42423 (* 0.0272727 = 0.0933881 loss) | |
| I0525 00:28:40.751333 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 3.3034 (* 0.0272727 = 0.0900927 loss) | |
| I0525 00:28:40.751346 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 2.48196 (* 0.0272727 = 0.0676897 loss) | |
| I0525 00:28:40.751360 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.81997 (* 0.0272727 = 0.0496355 loss) | |
| I0525 00:28:40.751374 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 2.00737 (* 0.0272727 = 0.0547465 loss) | |
| I0525 00:28:40.751387 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 1.77121 (* 0.0272727 = 0.0483058 loss) | |
| I0525 00:28:40.751400 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.753922 (* 0.0272727 = 0.0205615 loss) | |
| I0525 00:28:40.751420 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.710852 (* 0.0272727 = 0.0193869 loss) | |
| I0525 00:28:40.751435 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.854343 (* 0.0272727 = 0.0233003 loss) | |
| I0525 00:28:40.751447 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.851088 (* 0.0272727 = 0.0232115 loss) | |
| I0525 00:28:40.751462 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.038494 (* 0.0272727 = 0.00104984 loss) | |
| I0525 00:28:40.751476 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0161986 (* 0.0272727 = 0.000441781 loss) | |
| I0525 00:28:40.751489 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0204036 (* 0.0272727 = 0.000556463 loss) | |
| I0525 00:28:40.751503 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0299059 (* 0.0272727 = 0.000815615 loss) | |
| I0525 00:28:40.751516 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0284334 (* 0.0272727 = 0.000775455 loss) | |
| I0525 00:28:40.751530 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.013689 (* 0.0272727 = 0.000373336 loss) | |
| I0525 00:28:40.751544 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0209422 (* 0.0272727 = 0.000571152 loss) | |
| I0525 00:28:40.751555 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0166667 | |
| I0525 00:28:40.751567 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 00:28:40.751579 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:28:40.751590 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 00:28:40.751601 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 00:28:40.751612 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 00:28:40.751624 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 00:28:40.751636 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0525 00:28:40.751646 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0525 00:28:40.751658 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0525 00:28:40.751669 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0525 00:28:40.751682 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.75 | |
| I0525 00:28:40.751693 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 00:28:40.751703 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 00:28:40.751714 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 00:28:40.751725 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 00:28:40.751737 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 00:28:40.751759 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 00:28:40.751771 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 00:28:40.751782 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:28:40.751793 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:28:40.751807 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:28:40.751819 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:28:40.751832 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.664773 | |
| I0525 00:28:40.751847 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.116667 | |
| I0525 00:28:40.751858 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.13008 (* 1 = 4.13008 loss) | |
| I0525 00:28:40.751873 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.52206 (* 1 = 1.52206 loss) | |
| I0525 00:28:40.751885 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 4.16205 (* 0.0909091 = 0.378368 loss) | |
| I0525 00:28:40.751904 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.62899 (* 0.0909091 = 0.329908 loss) | |
| I0525 00:28:40.751919 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.92432 (* 0.0909091 = 0.356756 loss) | |
| I0525 00:28:40.751938 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.63787 (* 0.0909091 = 0.330716 loss) | |
| I0525 00:28:40.751952 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.51613 (* 0.0909091 = 0.319648 loss) | |
| I0525 00:28:40.751965 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.24685 (* 0.0909091 = 0.295169 loss) | |
| I0525 00:28:40.751979 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.84856 (* 0.0909091 = 0.25896 loss) | |
| I0525 00:28:40.751992 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.95696 (* 0.0909091 = 0.177905 loss) | |
| I0525 00:28:40.752005 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 1.51142 (* 0.0909091 = 0.137402 loss) | |
| I0525 00:28:40.752019 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 1.74986 (* 0.0909091 = 0.159078 loss) | |
| I0525 00:28:40.752032 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 1.61079 (* 0.0909091 = 0.146436 loss) | |
| I0525 00:28:40.752046 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.854904 (* 0.0909091 = 0.0777185 loss) | |
| I0525 00:28:40.752059 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 1.04248 (* 0.0909091 = 0.0947705 loss) | |
| I0525 00:28:40.752073 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.913568 (* 0.0909091 = 0.0830517 loss) | |
| I0525 00:28:40.752086 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.958453 (* 0.0909091 = 0.0871321 loss) | |
| I0525 00:28:40.752100 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0385833 (* 0.0909091 = 0.00350757 loss) | |
| I0525 00:28:40.752113 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0195122 (* 0.0909091 = 0.00177383 loss) | |
| I0525 00:28:40.752127 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.021645 (* 0.0909091 = 0.00196773 loss) | |
| I0525 00:28:40.752140 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00805381 (* 0.0909091 = 0.000732164 loss) | |
| I0525 00:28:40.752154 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0074033 (* 0.0909091 = 0.000673027 loss) | |
| I0525 00:28:40.752167 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00841593 (* 0.0909091 = 0.000765084 loss) | |
| I0525 00:28:40.752182 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00512228 (* 0.0909091 = 0.000465662 loss) | |
| I0525 00:28:40.752192 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:28:40.752203 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:28:40.752214 5272 solver.cpp:245] Train net output #149: total_confidence = 3.96097e-08 | |
| I0525 00:28:40.752235 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.2685e-06 | |
| I0525 00:28:40.752249 5272 sgd_solver.cpp:106] Iteration 1500, lr = 0.001 | |
| I0525 00:35:05.272518 5272 solver.cpp:229] Iteration 2000, loss = 12.7829 | |
| I0525 00:35:05.272686 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0227273 | |
| I0525 00:35:05.272707 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 00:35:05.272722 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0525 00:35:05.272733 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 00:35:05.272747 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 00:35:05.272759 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 00:35:05.272771 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 00:35:05.272783 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 00:35:05.272795 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 00:35:05.272807 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 00:35:05.272820 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 00:35:05.272831 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 00:35:05.272843 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 00:35:05.272855 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 00:35:05.272867 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 00:35:05.272881 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 00:35:05.272894 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 00:35:05.272907 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 00:35:05.272918 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 00:35:05.272929 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:35:05.272941 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:35:05.272953 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:35:05.272965 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:35:05.272976 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.755682 | |
| I0525 00:35:05.272989 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.159091 | |
| I0525 00:35:05.273005 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.92125 (* 0.3 = 1.17637 loss) | |
| I0525 00:35:05.273020 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.27123 (* 0.3 = 0.381368 loss) | |
| I0525 00:35:05.273035 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.53988 (* 0.0272727 = 0.0965421 loss) | |
| I0525 00:35:05.273048 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.48922 (* 0.0272727 = 0.0951607 loss) | |
| I0525 00:35:05.273062 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.59355 (* 0.0272727 = 0.0980059 loss) | |
| I0525 00:35:05.273077 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 4.47588 (* 0.0272727 = 0.122069 loss) | |
| I0525 00:35:05.273090 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 4.01635 (* 0.0272727 = 0.109537 loss) | |
| I0525 00:35:05.273104 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.89373 (* 0.0272727 = 0.0516471 loss) | |
| I0525 00:35:05.273133 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.07407 (* 0.0272727 = 0.0565656 loss) | |
| I0525 00:35:05.273150 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.04755 (* 0.0272727 = 0.0285696 loss) | |
| I0525 00:35:05.273164 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.126235 (* 0.0272727 = 0.00344278 loss) | |
| I0525 00:35:05.273180 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0787587 (* 0.0272727 = 0.00214796 loss) | |
| I0525 00:35:05.273193 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0573684 (* 0.0272727 = 0.00156459 loss) | |
| I0525 00:35:05.273208 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0460874 (* 0.0272727 = 0.00125693 loss) | |
| I0525 00:35:05.273222 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0604106 (* 0.0272727 = 0.00164756 loss) | |
| I0525 00:35:05.273257 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0426749 (* 0.0272727 = 0.00116386 loss) | |
| I0525 00:35:05.273273 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0395692 (* 0.0272727 = 0.00107916 loss) | |
| I0525 00:35:05.273288 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0445833 (* 0.0272727 = 0.00121591 loss) | |
| I0525 00:35:05.273301 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0271112 (* 0.0272727 = 0.000739395 loss) | |
| I0525 00:35:05.273315 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0140552 (* 0.0272727 = 0.000383324 loss) | |
| I0525 00:35:05.273329 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0313113 (* 0.0272727 = 0.000853944 loss) | |
| I0525 00:35:05.273344 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0358698 (* 0.0272727 = 0.000978267 loss) | |
| I0525 00:35:05.273357 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.014232 (* 0.0272727 = 0.000388146 loss) | |
| I0525 00:35:05.273371 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0281452 (* 0.0272727 = 0.000767597 loss) | |
| I0525 00:35:05.273385 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 00:35:05.273396 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 00:35:05.273408 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0525 00:35:05.273419 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 00:35:05.273432 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 00:35:05.273443 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 00:35:05.273455 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 00:35:05.273466 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 00:35:05.273478 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 00:35:05.273490 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 00:35:05.273501 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 00:35:05.273514 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 00:35:05.273524 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 00:35:05.273536 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 00:35:05.273547 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 00:35:05.273558 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 00:35:05.273571 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 00:35:05.273581 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 00:35:05.273593 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 00:35:05.273604 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:35:05.273617 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:35:05.273627 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:35:05.273639 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:35:05.273650 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0525 00:35:05.273663 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.181818 | |
| I0525 00:35:05.273676 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.93224 (* 0.3 = 1.17967 loss) | |
| I0525 00:35:05.273689 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.22689 (* 0.3 = 0.368068 loss) | |
| I0525 00:35:05.273707 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.7261 (* 0.0272727 = 0.101621 loss) | |
| I0525 00:35:05.273721 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.5749 (* 0.0272727 = 0.0974973 loss) | |
| I0525 00:35:05.273736 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.46198 (* 0.0272727 = 0.0944177 loss) | |
| I0525 00:35:05.273761 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 4.00962 (* 0.0272727 = 0.109353 loss) | |
| I0525 00:35:05.273775 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 4.0617 (* 0.0272727 = 0.110774 loss) | |
| I0525 00:35:05.273789 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.15543 (* 0.0272727 = 0.0587845 loss) | |
| I0525 00:35:05.273803 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.90413 (* 0.0272727 = 0.0519308 loss) | |
| I0525 00:35:05.273816 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.927545 (* 0.0272727 = 0.0252967 loss) | |
| I0525 00:35:05.273831 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.125155 (* 0.0272727 = 0.00341331 loss) | |
| I0525 00:35:05.273850 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0804279 (* 0.0272727 = 0.00219349 loss) | |
| I0525 00:35:05.273865 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0828219 (* 0.0272727 = 0.00225878 loss) | |
| I0525 00:35:05.273880 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0812313 (* 0.0272727 = 0.0022154 loss) | |
| I0525 00:35:05.273895 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0752477 (* 0.0272727 = 0.00205221 loss) | |
| I0525 00:35:05.273908 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.075886 (* 0.0272727 = 0.00206962 loss) | |
| I0525 00:35:05.273922 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0441645 (* 0.0272727 = 0.00120449 loss) | |
| I0525 00:35:05.273939 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.043339 (* 0.0272727 = 0.00118197 loss) | |
| I0525 00:35:05.273953 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0305005 (* 0.0272727 = 0.000831833 loss) | |
| I0525 00:35:05.273967 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0332818 (* 0.0272727 = 0.000907687 loss) | |
| I0525 00:35:05.273982 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0287694 (* 0.0272727 = 0.00078462 loss) | |
| I0525 00:35:05.273995 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0178319 (* 0.0272727 = 0.000486324 loss) | |
| I0525 00:35:05.274010 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0399262 (* 0.0272727 = 0.0010889 loss) | |
| I0525 00:35:05.274024 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0279846 (* 0.0272727 = 0.000763216 loss) | |
| I0525 00:35:05.274036 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0454545 | |
| I0525 00:35:05.274049 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 00:35:05.274060 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:35:05.274071 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 00:35:05.274083 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 00:35:05.274094 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 00:35:05.274106 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 00:35:05.274118 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 00:35:05.274130 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 00:35:05.274142 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 00:35:05.274153 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 00:35:05.274165 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 00:35:05.274176 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 00:35:05.274188 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 00:35:05.274199 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 00:35:05.274210 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 00:35:05.274222 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 00:35:05.274245 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 00:35:05.274257 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 00:35:05.274268 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:35:05.274281 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:35:05.274292 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:35:05.274304 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:35:05.274315 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.75 | |
| I0525 00:35:05.274327 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.159091 | |
| I0525 00:35:05.274341 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.75894 (* 1 = 3.75894 loss) | |
| I0525 00:35:05.274355 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.19701 (* 1 = 1.19701 loss) | |
| I0525 00:35:05.274369 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.54954 (* 0.0909091 = 0.322686 loss) | |
| I0525 00:35:05.274382 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.75869 (* 0.0909091 = 0.341699 loss) | |
| I0525 00:35:05.274396 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.3944 (* 0.0909091 = 0.308582 loss) | |
| I0525 00:35:05.274410 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.8402 (* 0.0909091 = 0.349109 loss) | |
| I0525 00:35:05.274425 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.65981 (* 0.0909091 = 0.33271 loss) | |
| I0525 00:35:05.274437 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.05177 (* 0.0909091 = 0.186524 loss) | |
| I0525 00:35:05.274451 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.69134 (* 0.0909091 = 0.153758 loss) | |
| I0525 00:35:05.274466 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.904634 (* 0.0909091 = 0.0822395 loss) | |
| I0525 00:35:05.274478 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.120506 (* 0.0909091 = 0.0109551 loss) | |
| I0525 00:35:05.274492 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0811063 (* 0.0909091 = 0.0073733 loss) | |
| I0525 00:35:05.274507 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0766133 (* 0.0909091 = 0.00696485 loss) | |
| I0525 00:35:05.274520 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0584315 (* 0.0909091 = 0.00531196 loss) | |
| I0525 00:35:05.274534 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0580073 (* 0.0909091 = 0.00527339 loss) | |
| I0525 00:35:05.274549 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0332099 (* 0.0909091 = 0.00301908 loss) | |
| I0525 00:35:05.274562 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0254822 (* 0.0909091 = 0.00231657 loss) | |
| I0525 00:35:05.274576 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0171924 (* 0.0909091 = 0.00156295 loss) | |
| I0525 00:35:05.274590 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00972325 (* 0.0909091 = 0.000883932 loss) | |
| I0525 00:35:05.274603 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00888911 (* 0.0909091 = 0.000808101 loss) | |
| I0525 00:35:05.274617 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0058786 (* 0.0909091 = 0.000534418 loss) | |
| I0525 00:35:05.274631 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0072391 (* 0.0909091 = 0.0006581 loss) | |
| I0525 00:35:05.274646 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00436002 (* 0.0909091 = 0.000396366 loss) | |
| I0525 00:35:05.274659 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00509385 (* 0.0909091 = 0.000463077 loss) | |
| I0525 00:35:05.274672 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:35:05.274683 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:35:05.274691 5272 solver.cpp:245] Train net output #149: total_confidence = 2.32681e-09 | |
| I0525 00:35:05.274713 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.42269e-06 | |
| I0525 00:35:05.274729 5272 sgd_solver.cpp:106] Iteration 2000, lr = 0.001 | |
| I0525 00:41:29.941377 5272 solver.cpp:229] Iteration 2500, loss = 12.789 | |
| I0525 00:41:29.941529 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0232558 | |
| I0525 00:41:29.941550 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 00:41:29.941563 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 00:41:29.941576 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 00:41:29.941587 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 00:41:29.941598 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 00:41:29.941612 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 00:41:29.941623 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 00:41:29.941637 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 00:41:29.941648 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 00:41:29.941659 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 00:41:29.941671 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 00:41:29.941684 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 00:41:29.941695 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 00:41:29.941707 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 00:41:29.941720 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 00:41:29.941732 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 00:41:29.941745 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 00:41:29.941756 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 00:41:29.941767 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:41:29.941781 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:41:29.941792 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:41:29.941803 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:41:29.941815 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.761364 | |
| I0525 00:41:29.941828 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0930233 | |
| I0525 00:41:29.941843 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.09851 (* 0.3 = 1.22955 loss) | |
| I0525 00:41:29.941859 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.19809 (* 0.3 = 0.359426 loss) | |
| I0525 00:41:29.941872 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.88507 (* 0.0272727 = 0.105956 loss) | |
| I0525 00:41:29.941890 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.89771 (* 0.0272727 = 0.106301 loss) | |
| I0525 00:41:29.941905 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.68213 (* 0.0272727 = 0.127695 loss) | |
| I0525 00:41:29.941917 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.99585 (* 0.0272727 = 0.108978 loss) | |
| I0525 00:41:29.941931 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.90896 (* 0.0272727 = 0.106608 loss) | |
| I0525 00:41:29.941946 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.34393 (* 0.0272727 = 0.0639254 loss) | |
| I0525 00:41:29.941958 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.00309 (* 0.0272727 = 0.0273569 loss) | |
| I0525 00:41:29.941973 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.309577 (* 0.0272727 = 0.008443 loss) | |
| I0525 00:41:29.941987 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.145605 (* 0.0272727 = 0.00397103 loss) | |
| I0525 00:41:29.942001 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.158907 (* 0.0272727 = 0.00433383 loss) | |
| I0525 00:41:29.942016 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0976243 (* 0.0272727 = 0.00266248 loss) | |
| I0525 00:41:29.942030 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.139111 (* 0.0272727 = 0.00379393 loss) | |
| I0525 00:41:29.942044 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0755138 (* 0.0272727 = 0.00205947 loss) | |
| I0525 00:41:29.942078 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0851903 (* 0.0272727 = 0.00232337 loss) | |
| I0525 00:41:29.942093 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0767541 (* 0.0272727 = 0.00209329 loss) | |
| I0525 00:41:29.942108 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0435614 (* 0.0272727 = 0.00118804 loss) | |
| I0525 00:41:29.942122 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.07218 (* 0.0272727 = 0.00196855 loss) | |
| I0525 00:41:29.942137 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0291235 (* 0.0272727 = 0.000794276 loss) | |
| I0525 00:41:29.942150 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.039615 (* 0.0272727 = 0.00108041 loss) | |
| I0525 00:41:29.942164 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0332074 (* 0.0272727 = 0.000905655 loss) | |
| I0525 00:41:29.942178 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0357341 (* 0.0272727 = 0.000974567 loss) | |
| I0525 00:41:29.942193 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0290932 (* 0.0272727 = 0.00079345 loss) | |
| I0525 00:41:29.942205 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 00:41:29.942217 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 00:41:29.942229 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 00:41:29.942240 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 00:41:29.942251 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 00:41:29.942263 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 00:41:29.942275 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 00:41:29.942286 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 00:41:29.942298 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 00:41:29.942309 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 00:41:29.942322 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 00:41:29.942332 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 00:41:29.942344 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 00:41:29.942356 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 00:41:29.942368 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 00:41:29.942375 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 00:41:29.942384 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 00:41:29.942395 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 00:41:29.942407 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 00:41:29.942419 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:41:29.942430 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:41:29.942441 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:41:29.942452 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:41:29.942464 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0525 00:41:29.942476 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0930233 | |
| I0525 00:41:29.942489 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.13513 (* 0.3 = 1.24054 loss) | |
| I0525 00:41:29.942503 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.19938 (* 0.3 = 0.359815 loss) | |
| I0525 00:41:29.942517 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.41432 (* 0.0272727 = 0.12039 loss) | |
| I0525 00:41:29.942530 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.14429 (* 0.0272727 = 0.113026 loss) | |
| I0525 00:41:29.942548 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.26111 (* 0.0272727 = 0.116212 loss) | |
| I0525 00:41:29.942574 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 4.07413 (* 0.0272727 = 0.111113 loss) | |
| I0525 00:41:29.942587 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.92212 (* 0.0272727 = 0.106967 loss) | |
| I0525 00:41:29.942601 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.0807 (* 0.0272727 = 0.0567463 loss) | |
| I0525 00:41:29.942615 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.31967 (* 0.0272727 = 0.0359911 loss) | |
| I0525 00:41:29.942630 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.289858 (* 0.0272727 = 0.00790522 loss) | |
| I0525 00:41:29.942643 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.147687 (* 0.0272727 = 0.00402782 loss) | |
| I0525 00:41:29.942657 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.123971 (* 0.0272727 = 0.00338102 loss) | |
| I0525 00:41:29.942672 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.147 (* 0.0272727 = 0.00400909 loss) | |
| I0525 00:41:29.942684 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0952754 (* 0.0272727 = 0.00259842 loss) | |
| I0525 00:41:29.942698 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.105477 (* 0.0272727 = 0.00287665 loss) | |
| I0525 00:41:29.942713 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0624571 (* 0.0272727 = 0.00170338 loss) | |
| I0525 00:41:29.942726 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.127862 (* 0.0272727 = 0.00348716 loss) | |
| I0525 00:41:29.942740 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0815335 (* 0.0272727 = 0.00222364 loss) | |
| I0525 00:41:29.942754 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0439283 (* 0.0272727 = 0.00119804 loss) | |
| I0525 00:41:29.942767 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0585262 (* 0.0272727 = 0.00159617 loss) | |
| I0525 00:41:29.942781 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0785093 (* 0.0272727 = 0.00214116 loss) | |
| I0525 00:41:29.942795 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0581652 (* 0.0272727 = 0.00158632 loss) | |
| I0525 00:41:29.942809 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0680433 (* 0.0272727 = 0.00185573 loss) | |
| I0525 00:41:29.942823 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0286516 (* 0.0272727 = 0.000781408 loss) | |
| I0525 00:41:29.942836 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0525 00:41:29.942847 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 00:41:29.942859 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:41:29.942870 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 00:41:29.942883 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 00:41:29.942893 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 00:41:29.942905 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 00:41:29.942917 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 00:41:29.942931 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 00:41:29.942944 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 00:41:29.942955 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 00:41:29.942966 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 00:41:29.942978 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 00:41:29.942989 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 00:41:29.943001 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 00:41:29.943012 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 00:41:29.943023 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 00:41:29.943035 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 00:41:29.943055 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 00:41:29.943068 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:41:29.943080 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:41:29.943091 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:41:29.943104 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:41:29.943114 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0525 00:41:29.943126 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0465116 | |
| I0525 00:41:29.943140 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.91883 (* 1 = 3.91883 loss) | |
| I0525 00:41:29.943153 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.18991 (* 1 = 1.18991 loss) | |
| I0525 00:41:29.943167 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.8382 (* 0.0909091 = 0.348927 loss) | |
| I0525 00:41:29.943181 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.78463 (* 0.0909091 = 0.344057 loss) | |
| I0525 00:41:29.943195 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.10665 (* 0.0909091 = 0.373332 loss) | |
| I0525 00:41:29.943209 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.90022 (* 0.0909091 = 0.354565 loss) | |
| I0525 00:41:29.943223 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.87986 (* 0.0909091 = 0.352715 loss) | |
| I0525 00:41:29.943236 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.33136 (* 0.0909091 = 0.211942 loss) | |
| I0525 00:41:29.943250 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.991427 (* 0.0909091 = 0.0901297 loss) | |
| I0525 00:41:29.943264 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.253009 (* 0.0909091 = 0.0230008 loss) | |
| I0525 00:41:29.943279 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.166617 (* 0.0909091 = 0.015147 loss) | |
| I0525 00:41:29.943292 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.125126 (* 0.0909091 = 0.0113751 loss) | |
| I0525 00:41:29.943305 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0899374 (* 0.0909091 = 0.00817613 loss) | |
| I0525 00:41:29.943320 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0972507 (* 0.0909091 = 0.00884098 loss) | |
| I0525 00:41:29.943333 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0888271 (* 0.0909091 = 0.00807519 loss) | |
| I0525 00:41:29.943347 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0637309 (* 0.0909091 = 0.00579372 loss) | |
| I0525 00:41:29.943361 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0531034 (* 0.0909091 = 0.00482758 loss) | |
| I0525 00:41:29.943375 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0408775 (* 0.0909091 = 0.00371613 loss) | |
| I0525 00:41:29.943389 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0162169 (* 0.0909091 = 0.00147426 loss) | |
| I0525 00:41:29.943403 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00952436 (* 0.0909091 = 0.000865851 loss) | |
| I0525 00:41:29.943418 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00832199 (* 0.0909091 = 0.000756544 loss) | |
| I0525 00:41:29.943431 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00610849 (* 0.0909091 = 0.000555317 loss) | |
| I0525 00:41:29.943445 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00634939 (* 0.0909091 = 0.000577217 loss) | |
| I0525 00:41:29.943459 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00501891 (* 0.0909091 = 0.000456264 loss) | |
| I0525 00:41:29.943471 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:41:29.943483 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:41:29.943495 5272 solver.cpp:245] Train net output #149: total_confidence = 8.6717e-09 | |
| I0525 00:41:29.943506 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.53332e-07 | |
| I0525 00:41:29.943528 5272 sgd_solver.cpp:106] Iteration 2500, lr = 0.001 | |
| I0525 00:47:54.740547 5272 solver.cpp:229] Iteration 3000, loss = 12.6279 | |
| I0525 00:47:54.740674 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0149254 | |
| I0525 00:47:54.740695 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 00:47:54.740708 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 00:47:54.740720 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 00:47:54.740732 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0525 00:47:54.740744 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 00:47:54.740757 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0525 00:47:54.740770 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.25 | |
| I0525 00:47:54.740782 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.5 | |
| I0525 00:47:54.740794 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0525 00:47:54.740805 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0525 00:47:54.740818 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 00:47:54.740829 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 00:47:54.740841 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 00:47:54.740854 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 00:47:54.740865 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 00:47:54.740880 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 00:47:54.740891 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0525 00:47:54.740903 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0525 00:47:54.740916 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:47:54.740927 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:47:54.740938 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:47:54.740949 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:47:54.740962 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.625 | |
| I0525 00:47:54.740973 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0746269 | |
| I0525 00:47:54.740989 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.24946 (* 0.3 = 1.27484 loss) | |
| I0525 00:47:54.741003 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.73441 (* 0.3 = 0.520322 loss) | |
| I0525 00:47:54.741017 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.62352 (* 0.0272727 = 0.126096 loss) | |
| I0525 00:47:54.741034 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.98568 (* 0.0272727 = 0.1087 loss) | |
| I0525 00:47:54.741047 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.70204 (* 0.0272727 = 0.100965 loss) | |
| I0525 00:47:54.741060 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.37382 (* 0.0272727 = 0.0920132 loss) | |
| I0525 00:47:54.741075 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.37772 (* 0.0272727 = 0.0921196 loss) | |
| I0525 00:47:54.741088 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 4.35502 (* 0.0272727 = 0.118773 loss) | |
| I0525 00:47:54.741102 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 3.97541 (* 0.0272727 = 0.10842 loss) | |
| I0525 00:47:54.741116 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 2.78381 (* 0.0272727 = 0.0759221 loss) | |
| I0525 00:47:54.741145 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.38435 (* 0.0272727 = 0.037755 loss) | |
| I0525 00:47:54.741159 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 1.50124 (* 0.0272727 = 0.040943 loss) | |
| I0525 00:47:54.741173 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.82868 (* 0.0272727 = 0.0226004 loss) | |
| I0525 00:47:54.741188 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.902771 (* 0.0272727 = 0.024621 loss) | |
| I0525 00:47:54.741219 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.758071 (* 0.0272727 = 0.0206747 loss) | |
| I0525 00:47:54.741235 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.910692 (* 0.0272727 = 0.0248371 loss) | |
| I0525 00:47:54.741248 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.766525 (* 0.0272727 = 0.0209052 loss) | |
| I0525 00:47:54.741262 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.99003 (* 0.0272727 = 0.0270008 loss) | |
| I0525 00:47:54.741276 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.679101 (* 0.0272727 = 0.0185209 loss) | |
| I0525 00:47:54.741289 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.704541 (* 0.0272727 = 0.0192148 loss) | |
| I0525 00:47:54.741304 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0199109 (* 0.0272727 = 0.000543024 loss) | |
| I0525 00:47:54.741318 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0297044 (* 0.0272727 = 0.000810121 loss) | |
| I0525 00:47:54.741331 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0268899 (* 0.0272727 = 0.00073336 loss) | |
| I0525 00:47:54.741345 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0212802 (* 0.0272727 = 0.00058037 loss) | |
| I0525 00:47:54.741358 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0149254 | |
| I0525 00:47:54.741369 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 00:47:54.741381 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 00:47:54.741392 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 00:47:54.741403 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 00:47:54.741415 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 00:47:54.741426 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0525 00:47:54.741438 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0525 00:47:54.741449 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.5 | |
| I0525 00:47:54.741461 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0525 00:47:54.741473 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0525 00:47:54.741484 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 00:47:54.741497 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 00:47:54.741508 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 00:47:54.741518 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 00:47:54.741530 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 00:47:54.741541 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 00:47:54.741552 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0525 00:47:54.741564 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0525 00:47:54.741575 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:47:54.741586 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:47:54.741598 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:47:54.741610 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:47:54.741621 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.625 | |
| I0525 00:47:54.741631 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0746269 | |
| I0525 00:47:54.741646 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.21009 (* 0.3 = 1.26303 loss) | |
| I0525 00:47:54.741659 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.73829 (* 0.3 = 0.521488 loss) | |
| I0525 00:47:54.741672 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.31466 (* 0.0272727 = 0.117672 loss) | |
| I0525 00:47:54.741688 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.11623 (* 0.0272727 = 0.112261 loss) | |
| I0525 00:47:54.741714 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.33444 (* 0.0272727 = 0.118212 loss) | |
| I0525 00:47:54.741729 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 4.05153 (* 0.0272727 = 0.110496 loss) | |
| I0525 00:47:54.741744 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.58904 (* 0.0272727 = 0.0978829 loss) | |
| I0525 00:47:54.741757 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 4.40439 (* 0.0272727 = 0.12012 loss) | |
| I0525 00:47:54.741770 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 4.24856 (* 0.0272727 = 0.11587 loss) | |
| I0525 00:47:54.741783 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 2.52428 (* 0.0272727 = 0.0688441 loss) | |
| I0525 00:47:54.741797 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.38601 (* 0.0272727 = 0.0378002 loss) | |
| I0525 00:47:54.741811 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 1.62495 (* 0.0272727 = 0.0443169 loss) | |
| I0525 00:47:54.741824 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.757404 (* 0.0272727 = 0.0206565 loss) | |
| I0525 00:47:54.741838 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.728316 (* 0.0272727 = 0.0198632 loss) | |
| I0525 00:47:54.741855 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.813051 (* 0.0272727 = 0.0221741 loss) | |
| I0525 00:47:54.741865 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.766242 (* 0.0272727 = 0.0208975 loss) | |
| I0525 00:47:54.741879 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.945189 (* 0.0272727 = 0.0257779 loss) | |
| I0525 00:47:54.741894 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 1.07776 (* 0.0272727 = 0.0293935 loss) | |
| I0525 00:47:54.741907 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 1.0765 (* 0.0272727 = 0.029359 loss) | |
| I0525 00:47:54.741920 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 1.13546 (* 0.0272727 = 0.0309672 loss) | |
| I0525 00:47:54.741937 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.012686 (* 0.0272727 = 0.000345981 loss) | |
| I0525 00:47:54.741951 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0177177 (* 0.0272727 = 0.00048321 loss) | |
| I0525 00:47:54.741966 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0177239 (* 0.0272727 = 0.000483379 loss) | |
| I0525 00:47:54.741979 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0213743 (* 0.0272727 = 0.000582935 loss) | |
| I0525 00:47:54.741991 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0597015 | |
| I0525 00:47:54.742003 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 00:47:54.742015 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:47:54.742027 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 00:47:54.742038 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 00:47:54.742048 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 00:47:54.742060 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0525 00:47:54.742071 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.25 | |
| I0525 00:47:54.742082 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.5 | |
| I0525 00:47:54.742094 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0525 00:47:54.742105 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0525 00:47:54.742116 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 00:47:54.742127 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 00:47:54.742138 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 00:47:54.742149 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 00:47:54.742161 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 00:47:54.742172 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 00:47:54.742193 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0525 00:47:54.742207 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0525 00:47:54.742218 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:47:54.742229 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:47:54.742240 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:47:54.742251 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:47:54.742262 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.636364 | |
| I0525 00:47:54.742274 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.134328 | |
| I0525 00:47:54.742287 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.95572 (* 1 = 3.95572 loss) | |
| I0525 00:47:54.742300 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.63339 (* 1 = 1.63339 loss) | |
| I0525 00:47:54.742314 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 4.03244 (* 0.0909091 = 0.366586 loss) | |
| I0525 00:47:54.742327 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.88254 (* 0.0909091 = 0.352959 loss) | |
| I0525 00:47:54.742341 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.2057 (* 0.0909091 = 0.382336 loss) | |
| I0525 00:47:54.742354 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.75199 (* 0.0909091 = 0.34109 loss) | |
| I0525 00:47:54.742367 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.51788 (* 0.0909091 = 0.319807 loss) | |
| I0525 00:47:54.742382 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 4.16721 (* 0.0909091 = 0.378837 loss) | |
| I0525 00:47:54.742394 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 4.01801 (* 0.0909091 = 0.365273 loss) | |
| I0525 00:47:54.742408 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 2.59122 (* 0.0909091 = 0.235566 loss) | |
| I0525 00:47:54.742422 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 1.37916 (* 0.0909091 = 0.125379 loss) | |
| I0525 00:47:54.742436 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 1.68781 (* 0.0909091 = 0.153437 loss) | |
| I0525 00:47:54.742449 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.878813 (* 0.0909091 = 0.0798921 loss) | |
| I0525 00:47:54.742463 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.97025 (* 0.0909091 = 0.0882046 loss) | |
| I0525 00:47:54.742477 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.779983 (* 0.0909091 = 0.0709075 loss) | |
| I0525 00:47:54.742491 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.959548 (* 0.0909091 = 0.0872317 loss) | |
| I0525 00:47:54.742504 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 1.01181 (* 0.0909091 = 0.0919831 loss) | |
| I0525 00:47:54.742517 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.999385 (* 0.0909091 = 0.0908532 loss) | |
| I0525 00:47:54.742532 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 1.2285 (* 0.0909091 = 0.111681 loss) | |
| I0525 00:47:54.742544 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 1.30267 (* 0.0909091 = 0.118424 loss) | |
| I0525 00:47:54.742558 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00785119 (* 0.0909091 = 0.000713744 loss) | |
| I0525 00:47:54.742573 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00550414 (* 0.0909091 = 0.000500376 loss) | |
| I0525 00:47:54.742585 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00578178 (* 0.0909091 = 0.000525616 loss) | |
| I0525 00:47:54.742599 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00487508 (* 0.0909091 = 0.000443189 loss) | |
| I0525 00:47:54.742611 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:47:54.742624 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:47:54.742635 5272 solver.cpp:245] Train net output #149: total_confidence = 1.25812e-08 | |
| I0525 00:47:54.742655 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00010585 | |
| I0525 00:47:54.742671 5272 sgd_solver.cpp:106] Iteration 3000, lr = 0.001 | |
| I0525 00:51:00.725512 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.8013 > 30) by scale factor 0.887541 | |
| I0525 00:54:19.569465 5272 solver.cpp:229] Iteration 3500, loss = 12.3235 | |
| I0525 00:54:19.569600 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.04 | |
| I0525 00:54:19.569631 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 00:54:19.569656 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 00:54:19.569679 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 00:54:19.569700 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 00:54:19.569723 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 00:54:19.569746 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 00:54:19.569771 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 00:54:19.569792 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 00:54:19.569815 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 00:54:19.569836 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 00:54:19.569857 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 00:54:19.569882 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 00:54:19.569905 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 00:54:19.569926 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 00:54:19.569947 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 00:54:19.569967 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 00:54:19.569989 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 00:54:19.570009 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 00:54:19.570029 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 00:54:19.570051 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 00:54:19.570072 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 00:54:19.570092 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 00:54:19.570112 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.727273 | |
| I0525 00:54:19.570133 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.06 | |
| I0525 00:54:19.570160 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 4.05289 (* 0.3 = 1.21587 loss) | |
| I0525 00:54:19.570186 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.30261 (* 0.3 = 0.390783 loss) | |
| I0525 00:54:19.570211 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.6169 (* 0.0272727 = 0.0986427 loss) | |
| I0525 00:54:19.570240 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.12462 (* 0.0272727 = 0.11249 loss) | |
| I0525 00:54:19.570269 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.2791 (* 0.0272727 = 0.116703 loss) | |
| I0525 00:54:19.570297 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.66772 (* 0.0272727 = 0.100029 loss) | |
| I0525 00:54:19.570322 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.93488 (* 0.0272727 = 0.107315 loss) | |
| I0525 00:54:19.570346 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.64059 (* 0.0272727 = 0.0992888 loss) | |
| I0525 00:54:19.570371 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.60979 (* 0.0272727 = 0.071176 loss) | |
| I0525 00:54:19.570397 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.766985 (* 0.0272727 = 0.0209178 loss) | |
| I0525 00:54:19.570423 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.686941 (* 0.0272727 = 0.0187348 loss) | |
| I0525 00:54:19.570451 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.174644 (* 0.0272727 = 0.00476303 loss) | |
| I0525 00:54:19.570477 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.128736 (* 0.0272727 = 0.00351097 loss) | |
| I0525 00:54:19.570502 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0868422 (* 0.0272727 = 0.00236842 loss) | |
| I0525 00:54:19.570528 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0528018 (* 0.0272727 = 0.00144005 loss) | |
| I0525 00:54:19.570577 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0887583 (* 0.0272727 = 0.00242068 loss) | |
| I0525 00:54:19.570605 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0665068 (* 0.0272727 = 0.00181382 loss) | |
| I0525 00:54:19.570639 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0624142 (* 0.0272727 = 0.0017022 loss) | |
| I0525 00:54:19.570667 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0258059 (* 0.0272727 = 0.000703796 loss) | |
| I0525 00:54:19.570693 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0243421 (* 0.0272727 = 0.000663875 loss) | |
| I0525 00:54:19.570718 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0162986 (* 0.0272727 = 0.000444507 loss) | |
| I0525 00:54:19.570744 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0347205 (* 0.0272727 = 0.000946922 loss) | |
| I0525 00:54:19.570770 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0296229 (* 0.0272727 = 0.000807897 loss) | |
| I0525 00:54:19.570794 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0319445 (* 0.0272727 = 0.000871213 loss) | |
| I0525 00:54:19.570816 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.02 | |
| I0525 00:54:19.570838 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 00:54:19.570859 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 00:54:19.570879 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 00:54:19.570897 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 00:54:19.570914 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 00:54:19.570936 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 00:54:19.570958 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 00:54:19.570979 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 00:54:19.571001 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 00:54:19.571022 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 00:54:19.571043 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 00:54:19.571065 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 00:54:19.571085 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 00:54:19.571106 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 00:54:19.571126 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 00:54:19.571147 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 00:54:19.571167 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 00:54:19.571188 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 00:54:19.571208 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 00:54:19.571229 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 00:54:19.571250 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 00:54:19.571271 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 00:54:19.571291 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.721591 | |
| I0525 00:54:19.571312 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.08 | |
| I0525 00:54:19.571338 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.01737 (* 0.3 = 1.20521 loss) | |
| I0525 00:54:19.571363 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.31319 (* 0.3 = 0.393957 loss) | |
| I0525 00:54:19.571388 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.72539 (* 0.0272727 = 0.101601 loss) | |
| I0525 00:54:19.571413 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.90458 (* 0.0272727 = 0.106488 loss) | |
| I0525 00:54:19.571439 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.10439 (* 0.0272727 = 0.111938 loss) | |
| I0525 00:54:19.571481 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.39497 (* 0.0272727 = 0.0925902 loss) | |
| I0525 00:54:19.571507 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.85229 (* 0.0272727 = 0.105062 loss) | |
| I0525 00:54:19.571532 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.55364 (* 0.0272727 = 0.0969174 loss) | |
| I0525 00:54:19.571557 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.68186 (* 0.0272727 = 0.0731416 loss) | |
| I0525 00:54:19.571583 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.830551 (* 0.0272727 = 0.0226514 loss) | |
| I0525 00:54:19.571607 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.859805 (* 0.0272727 = 0.0234492 loss) | |
| I0525 00:54:19.571633 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.337482 (* 0.0272727 = 0.00920407 loss) | |
| I0525 00:54:19.571665 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.148599 (* 0.0272727 = 0.0040527 loss) | |
| I0525 00:54:19.571692 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.219055 (* 0.0272727 = 0.00597422 loss) | |
| I0525 00:54:19.571717 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.192707 (* 0.0272727 = 0.00525566 loss) | |
| I0525 00:54:19.571743 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.116718 (* 0.0272727 = 0.00318322 loss) | |
| I0525 00:54:19.571769 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.127619 (* 0.0272727 = 0.00348052 loss) | |
| I0525 00:54:19.571794 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.124627 (* 0.0272727 = 0.00339893 loss) | |
| I0525 00:54:19.571820 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.114652 (* 0.0272727 = 0.00312687 loss) | |
| I0525 00:54:19.571846 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.135244 (* 0.0272727 = 0.00368848 loss) | |
| I0525 00:54:19.571871 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0786351 (* 0.0272727 = 0.00214459 loss) | |
| I0525 00:54:19.571897 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0842862 (* 0.0272727 = 0.00229871 loss) | |
| I0525 00:54:19.571923 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.120936 (* 0.0272727 = 0.00329824 loss) | |
| I0525 00:54:19.571945 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0646458 (* 0.0272727 = 0.00176307 loss) | |
| I0525 00:54:19.571970 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.06 | |
| I0525 00:54:19.571997 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 00:54:19.572021 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 00:54:19.572041 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 00:54:19.572062 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 00:54:19.572083 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 00:54:19.572104 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 00:54:19.572125 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 00:54:19.572146 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 00:54:19.572167 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 00:54:19.572187 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 00:54:19.572209 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 00:54:19.572229 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 00:54:19.572250 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 00:54:19.572271 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 00:54:19.572291 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 00:54:19.572312 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 00:54:19.572332 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 00:54:19.572371 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 00:54:19.572393 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 00:54:19.572414 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 00:54:19.572435 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 00:54:19.572455 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 00:54:19.572476 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0525 00:54:19.572499 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.18 | |
| I0525 00:54:19.572525 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.81453 (* 1 = 3.81453 loss) | |
| I0525 00:54:19.572548 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.17502 (* 1 = 1.17502 loss) | |
| I0525 00:54:19.572573 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.54513 (* 0.0909091 = 0.322285 loss) | |
| I0525 00:54:19.572599 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 4.13379 (* 0.0909091 = 0.375799 loss) | |
| I0525 00:54:19.572624 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.02421 (* 0.0909091 = 0.365837 loss) | |
| I0525 00:54:19.572649 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.46215 (* 0.0909091 = 0.314741 loss) | |
| I0525 00:54:19.572675 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.68126 (* 0.0909091 = 0.33466 loss) | |
| I0525 00:54:19.572700 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.33668 (* 0.0909091 = 0.303335 loss) | |
| I0525 00:54:19.572734 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.37313 (* 0.0909091 = 0.215739 loss) | |
| I0525 00:54:19.572762 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.62296 (* 0.0909091 = 0.0566327 loss) | |
| I0525 00:54:19.572787 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.802836 (* 0.0909091 = 0.0729851 loss) | |
| I0525 00:54:19.572813 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.105153 (* 0.0909091 = 0.00955941 loss) | |
| I0525 00:54:19.572839 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.08439 (* 0.0909091 = 0.00767182 loss) | |
| I0525 00:54:19.572865 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0541126 (* 0.0909091 = 0.00491933 loss) | |
| I0525 00:54:19.572890 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0473339 (* 0.0909091 = 0.00430308 loss) | |
| I0525 00:54:19.572914 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.037998 (* 0.0909091 = 0.00345436 loss) | |
| I0525 00:54:19.572942 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0308387 (* 0.0909091 = 0.00280352 loss) | |
| I0525 00:54:19.572967 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0181048 (* 0.0909091 = 0.00164589 loss) | |
| I0525 00:54:19.572991 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0133417 (* 0.0909091 = 0.00121288 loss) | |
| I0525 00:54:19.573017 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00768539 (* 0.0909091 = 0.000698672 loss) | |
| I0525 00:54:19.573047 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00717772 (* 0.0909091 = 0.00065252 loss) | |
| I0525 00:54:19.573073 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00418342 (* 0.0909091 = 0.000380311 loss) | |
| I0525 00:54:19.573101 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00446216 (* 0.0909091 = 0.000405651 loss) | |
| I0525 00:54:19.573143 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0033466 (* 0.0909091 = 0.000304236 loss) | |
| I0525 00:54:19.573170 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 00:54:19.573192 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 00:54:19.573213 5272 solver.cpp:245] Train net output #149: total_confidence = 1.8921e-06 | |
| I0525 00:54:19.573235 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.47078e-05 | |
| I0525 00:54:19.573276 5272 sgd_solver.cpp:106] Iteration 3500, lr = 0.001 | |
| I0525 01:00:44.207233 5272 solver.cpp:229] Iteration 4000, loss = 12.2027 | |
| I0525 01:00:44.207389 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0784314 | |
| I0525 01:00:44.207409 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 01:00:44.207422 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 01:00:44.207434 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 01:00:44.207447 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 01:00:44.207458 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0525 01:00:44.207470 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 01:00:44.207482 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 01:00:44.207494 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 01:00:44.207507 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 01:00:44.207520 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 01:00:44.207531 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 01:00:44.207543 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:00:44.207556 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:00:44.207567 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:00:44.207578 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:00:44.207590 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:00:44.207602 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:00:44.207613 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:00:44.207625 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:00:44.207638 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:00:44.207649 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:00:44.207660 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:00:44.207672 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0525 01:00:44.207685 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.235294 | |
| I0525 01:00:44.207705 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.7898 (* 0.3 = 1.13694 loss) | |
| I0525 01:00:44.207720 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.40071 (* 0.3 = 0.420214 loss) | |
| I0525 01:00:44.207734 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.56248 (* 0.0272727 = 0.0971586 loss) | |
| I0525 01:00:44.207748 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.15332 (* 0.0272727 = 0.113272 loss) | |
| I0525 01:00:44.207762 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.1731 (* 0.0272727 = 0.113812 loss) | |
| I0525 01:00:44.207777 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 4.2464 (* 0.0272727 = 0.115811 loss) | |
| I0525 01:00:44.207790 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 4.54459 (* 0.0272727 = 0.123943 loss) | |
| I0525 01:00:44.207804 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.27699 (* 0.0272727 = 0.0893724 loss) | |
| I0525 01:00:44.207818 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.56313 (* 0.0272727 = 0.0426307 loss) | |
| I0525 01:00:44.207833 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.40193 (* 0.0272727 = 0.0382346 loss) | |
| I0525 01:00:44.207846 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.924763 (* 0.0272727 = 0.0252208 loss) | |
| I0525 01:00:44.207860 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.940588 (* 0.0272727 = 0.0256524 loss) | |
| I0525 01:00:44.207875 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.113318 (* 0.0272727 = 0.0030905 loss) | |
| I0525 01:00:44.207890 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0450705 (* 0.0272727 = 0.0012292 loss) | |
| I0525 01:00:44.207904 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.027961 (* 0.0272727 = 0.000762574 loss) | |
| I0525 01:00:44.207940 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0386235 (* 0.0272727 = 0.00105337 loss) | |
| I0525 01:00:44.207955 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0196546 (* 0.0272727 = 0.000536034 loss) | |
| I0525 01:00:44.207969 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0193924 (* 0.0272727 = 0.000528883 loss) | |
| I0525 01:00:44.207984 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0110779 (* 0.0272727 = 0.000302125 loss) | |
| I0525 01:00:44.207998 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0192586 (* 0.0272727 = 0.000525235 loss) | |
| I0525 01:00:44.208012 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0179756 (* 0.0272727 = 0.000490245 loss) | |
| I0525 01:00:44.208026 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00727792 (* 0.0272727 = 0.000198489 loss) | |
| I0525 01:00:44.208040 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00470918 (* 0.0272727 = 0.000128432 loss) | |
| I0525 01:00:44.208055 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0066181 (* 0.0272727 = 0.000180494 loss) | |
| I0525 01:00:44.208067 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0588235 | |
| I0525 01:00:44.208079 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 01:00:44.208091 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 01:00:44.208103 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:00:44.208115 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 01:00:44.208127 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0525 01:00:44.208138 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 01:00:44.208149 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 01:00:44.208161 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 01:00:44.208173 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 01:00:44.208186 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 01:00:44.208199 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 01:00:44.208209 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:00:44.208221 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:00:44.208233 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:00:44.208245 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:00:44.208256 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:00:44.208267 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:00:44.208278 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:00:44.208289 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:00:44.208300 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:00:44.208312 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:00:44.208324 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:00:44.208335 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.721591 | |
| I0525 01:00:44.208348 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0980392 | |
| I0525 01:00:44.208361 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.72966 (* 0.3 = 1.1189 loss) | |
| I0525 01:00:44.208375 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.41466 (* 0.3 = 0.424399 loss) | |
| I0525 01:00:44.208392 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.97659 (* 0.0272727 = 0.108452 loss) | |
| I0525 01:00:44.208406 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.44373 (* 0.0272727 = 0.09392 loss) | |
| I0525 01:00:44.208432 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.38585 (* 0.0272727 = 0.119614 loss) | |
| I0525 01:00:44.208446 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.97046 (* 0.0272727 = 0.108285 loss) | |
| I0525 01:00:44.208461 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 4.13731 (* 0.0272727 = 0.112836 loss) | |
| I0525 01:00:44.208474 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.48074 (* 0.0272727 = 0.0949293 loss) | |
| I0525 01:00:44.208488 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.64627 (* 0.0272727 = 0.0448984 loss) | |
| I0525 01:00:44.208501 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.62735 (* 0.0272727 = 0.0443823 loss) | |
| I0525 01:00:44.208515 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.684521 (* 0.0272727 = 0.0186688 loss) | |
| I0525 01:00:44.208529 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 1.07091 (* 0.0272727 = 0.0292067 loss) | |
| I0525 01:00:44.208544 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.103075 (* 0.0272727 = 0.00281115 loss) | |
| I0525 01:00:44.208559 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0655288 (* 0.0272727 = 0.00178715 loss) | |
| I0525 01:00:44.208572 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0295464 (* 0.0272727 = 0.000805811 loss) | |
| I0525 01:00:44.208586 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0470807 (* 0.0272727 = 0.00128402 loss) | |
| I0525 01:00:44.208600 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0281291 (* 0.0272727 = 0.000767158 loss) | |
| I0525 01:00:44.208616 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0386638 (* 0.0272727 = 0.00105447 loss) | |
| I0525 01:00:44.208629 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0501453 (* 0.0272727 = 0.0013676 loss) | |
| I0525 01:00:44.208644 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0227461 (* 0.0272727 = 0.000620349 loss) | |
| I0525 01:00:44.208658 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0133243 (* 0.0272727 = 0.000363389 loss) | |
| I0525 01:00:44.208673 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0135677 (* 0.0272727 = 0.000370029 loss) | |
| I0525 01:00:44.208685 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0372969 (* 0.0272727 = 0.00101719 loss) | |
| I0525 01:00:44.208699 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0168166 (* 0.0272727 = 0.000458634 loss) | |
| I0525 01:00:44.208712 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0392157 | |
| I0525 01:00:44.208724 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 01:00:44.208735 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:00:44.208750 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:00:44.208762 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 01:00:44.208775 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0525 01:00:44.208783 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 01:00:44.208791 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 01:00:44.208798 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 01:00:44.208811 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 01:00:44.208822 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 01:00:44.208834 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 01:00:44.208847 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:00:44.208858 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:00:44.208868 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:00:44.208879 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:00:44.208891 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:00:44.208912 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:00:44.208925 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:00:44.208937 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:00:44.208950 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:00:44.208961 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:00:44.208972 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:00:44.208983 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0525 01:00:44.208997 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.0980392 | |
| I0525 01:00:44.209010 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.61695 (* 1 = 3.61695 loss) | |
| I0525 01:00:44.209023 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.46096 (* 1 = 1.46096 loss) | |
| I0525 01:00:44.209038 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.31787 (* 0.0909091 = 0.301625 loss) | |
| I0525 01:00:44.209051 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.79495 (* 0.0909091 = 0.344995 loss) | |
| I0525 01:00:44.209064 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.81922 (* 0.0909091 = 0.347202 loss) | |
| I0525 01:00:44.209079 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.78443 (* 0.0909091 = 0.344039 loss) | |
| I0525 01:00:44.209092 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 4.17981 (* 0.0909091 = 0.379983 loss) | |
| I0525 01:00:44.209105 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.38597 (* 0.0909091 = 0.307816 loss) | |
| I0525 01:00:44.209133 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.61152 (* 0.0909091 = 0.146502 loss) | |
| I0525 01:00:44.209151 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.03527 (* 0.0909091 = 0.0941153 loss) | |
| I0525 01:00:44.209164 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.832669 (* 0.0909091 = 0.0756972 loss) | |
| I0525 01:00:44.209177 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.91665 (* 0.0909091 = 0.0833318 loss) | |
| I0525 01:00:44.209192 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.102732 (* 0.0909091 = 0.00933928 loss) | |
| I0525 01:00:44.209206 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.119427 (* 0.0909091 = 0.010857 loss) | |
| I0525 01:00:44.209219 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0545666 (* 0.0909091 = 0.0049606 loss) | |
| I0525 01:00:44.209233 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0432863 (* 0.0909091 = 0.00393512 loss) | |
| I0525 01:00:44.209247 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0279589 (* 0.0909091 = 0.00254171 loss) | |
| I0525 01:00:44.209261 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0160858 (* 0.0909091 = 0.00146235 loss) | |
| I0525 01:00:44.209275 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00682722 (* 0.0909091 = 0.000620656 loss) | |
| I0525 01:00:44.209288 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0033138 (* 0.0909091 = 0.000301255 loss) | |
| I0525 01:00:44.209303 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00289934 (* 0.0909091 = 0.000263576 loss) | |
| I0525 01:00:44.209317 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00295362 (* 0.0909091 = 0.000268511 loss) | |
| I0525 01:00:44.209331 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00258164 (* 0.0909091 = 0.000234694 loss) | |
| I0525 01:00:44.209345 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00228437 (* 0.0909091 = 0.00020767 loss) | |
| I0525 01:00:44.209358 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:00:44.209369 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:00:44.209381 5272 solver.cpp:245] Train net output #149: total_confidence = 4.48232e-09 | |
| I0525 01:00:44.209403 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 4.0606e-07 | |
| I0525 01:00:44.209419 5272 sgd_solver.cpp:106] Iteration 4000, lr = 0.001 | |
| I0525 01:06:38.358430 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.2945 > 30) by scale factor 0.84999 | |
| I0525 01:06:41.443186 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.4866 > 30) by scale factor 0.740985 | |
| I0525 01:07:08.783363 5272 solver.cpp:229] Iteration 4500, loss = 11.9827 | |
| I0525 01:07:08.783520 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0208333 | |
| I0525 01:07:08.783541 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 01:07:08.783555 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 01:07:08.783566 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 01:07:08.783578 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 01:07:08.783591 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 01:07:08.783602 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 01:07:08.783615 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 01:07:08.783627 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 01:07:08.783641 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 01:07:08.783653 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 01:07:08.783665 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 01:07:08.783677 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 01:07:08.783690 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 01:07:08.783702 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 01:07:08.783715 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:07:08.783726 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:07:08.783738 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:07:08.783751 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:07:08.783762 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:07:08.783774 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:07:08.783787 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:07:08.783798 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:07:08.783809 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.715909 | |
| I0525 01:07:08.783821 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.125 | |
| I0525 01:07:08.783838 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.79793 (* 0.3 = 1.13938 loss) | |
| I0525 01:07:08.783852 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.27764 (* 0.3 = 0.383293 loss) | |
| I0525 01:07:08.783867 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.29758 (* 0.0272727 = 0.117207 loss) | |
| I0525 01:07:08.783885 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.08246 (* 0.0272727 = 0.11134 loss) | |
| I0525 01:07:08.783900 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.77282 (* 0.0272727 = 0.102895 loss) | |
| I0525 01:07:08.783913 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.5288 (* 0.0272727 = 0.09624 loss) | |
| I0525 01:07:08.783928 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.54183 (* 0.0272727 = 0.0965954 loss) | |
| I0525 01:07:08.783942 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.30282 (* 0.0272727 = 0.0628042 loss) | |
| I0525 01:07:08.783957 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.62065 (* 0.0272727 = 0.0441996 loss) | |
| I0525 01:07:08.783970 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.01805 (* 0.0272727 = 0.027765 loss) | |
| I0525 01:07:08.783984 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.969495 (* 0.0272727 = 0.0264408 loss) | |
| I0525 01:07:08.783998 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 1.06902 (* 0.0272727 = 0.029155 loss) | |
| I0525 01:07:08.784013 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 1.04953 (* 0.0272727 = 0.0286235 loss) | |
| I0525 01:07:08.784026 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 1.37767 (* 0.0272727 = 0.0375727 loss) | |
| I0525 01:07:08.784041 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 1.28079 (* 0.0272727 = 0.0349306 loss) | |
| I0525 01:07:08.784077 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 1.14559 (* 0.0272727 = 0.0312433 loss) | |
| I0525 01:07:08.784093 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0276929 (* 0.0272727 = 0.00075526 loss) | |
| I0525 01:07:08.784107 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0252679 (* 0.0272727 = 0.000689125 loss) | |
| I0525 01:07:08.784122 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0147689 (* 0.0272727 = 0.000402788 loss) | |
| I0525 01:07:08.784137 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0118867 (* 0.0272727 = 0.000324184 loss) | |
| I0525 01:07:08.784150 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0108161 (* 0.0272727 = 0.000294985 loss) | |
| I0525 01:07:08.784164 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0129691 (* 0.0272727 = 0.000353704 loss) | |
| I0525 01:07:08.784178 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0119833 (* 0.0272727 = 0.000326817 loss) | |
| I0525 01:07:08.784193 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0096225 (* 0.0272727 = 0.000262432 loss) | |
| I0525 01:07:08.784205 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 01:07:08.784217 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 01:07:08.784229 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 01:07:08.784240 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:07:08.784251 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 01:07:08.784263 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 01:07:08.784276 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 01:07:08.784287 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 01:07:08.784301 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 01:07:08.784312 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 01:07:08.784324 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 01:07:08.784337 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 01:07:08.784348 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 01:07:08.784360 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 01:07:08.784373 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 01:07:08.784384 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:07:08.784395 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:07:08.784407 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:07:08.784418 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:07:08.784430 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:07:08.784441 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:07:08.784453 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:07:08.784464 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:07:08.784476 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0525 01:07:08.784487 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.0625 | |
| I0525 01:07:08.784502 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.05831 (* 0.3 = 1.21749 loss) | |
| I0525 01:07:08.784515 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.32194 (* 0.3 = 0.396583 loss) | |
| I0525 01:07:08.784533 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.29454 (* 0.0272727 = 0.117124 loss) | |
| I0525 01:07:08.784548 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.29894 (* 0.0272727 = 0.117244 loss) | |
| I0525 01:07:08.784571 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.75448 (* 0.0272727 = 0.102395 loss) | |
| I0525 01:07:08.784586 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.9861 (* 0.0272727 = 0.108712 loss) | |
| I0525 01:07:08.784600 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.36769 (* 0.0272727 = 0.0918461 loss) | |
| I0525 01:07:08.784615 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.19088 (* 0.0272727 = 0.0597512 loss) | |
| I0525 01:07:08.784628 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.56729 (* 0.0272727 = 0.0427443 loss) | |
| I0525 01:07:08.784642 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.960175 (* 0.0272727 = 0.0261866 loss) | |
| I0525 01:07:08.784657 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.918177 (* 0.0272727 = 0.0250412 loss) | |
| I0525 01:07:08.784670 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.955058 (* 0.0272727 = 0.026047 loss) | |
| I0525 01:07:08.784684 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.794209 (* 0.0272727 = 0.0216602 loss) | |
| I0525 01:07:08.784698 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 1.16649 (* 0.0272727 = 0.0318134 loss) | |
| I0525 01:07:08.784713 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.886968 (* 0.0272727 = 0.02419 loss) | |
| I0525 01:07:08.784726 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.82919 (* 0.0272727 = 0.0226143 loss) | |
| I0525 01:07:08.784740 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.062004 (* 0.0272727 = 0.00169102 loss) | |
| I0525 01:07:08.784754 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0467766 (* 0.0272727 = 0.00127573 loss) | |
| I0525 01:07:08.784768 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0457334 (* 0.0272727 = 0.00124728 loss) | |
| I0525 01:07:08.784782 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0400594 (* 0.0272727 = 0.00109253 loss) | |
| I0525 01:07:08.784796 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0234523 (* 0.0272727 = 0.000639608 loss) | |
| I0525 01:07:08.784811 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0289231 (* 0.0272727 = 0.000788811 loss) | |
| I0525 01:07:08.784824 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0274025 (* 0.0272727 = 0.000747342 loss) | |
| I0525 01:07:08.784837 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0184842 (* 0.0272727 = 0.000504115 loss) | |
| I0525 01:07:08.784850 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0 | |
| I0525 01:07:08.784862 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 01:07:08.784873 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 01:07:08.784885 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:07:08.784898 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 01:07:08.784905 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 01:07:08.784914 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 01:07:08.784929 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 01:07:08.784941 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 01:07:08.784952 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 01:07:08.784965 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 01:07:08.784976 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 01:07:08.784988 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 01:07:08.784999 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 01:07:08.785012 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 01:07:08.785022 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:07:08.785034 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:07:08.785055 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:07:08.785068 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:07:08.785080 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:07:08.785092 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:07:08.785104 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:07:08.785115 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:07:08.785145 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.721591 | |
| I0525 01:07:08.785157 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.166667 | |
| I0525 01:07:08.785171 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.1799 (* 1 = 4.1799 loss) | |
| I0525 01:07:08.785186 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.36976 (* 1 = 1.36976 loss) | |
| I0525 01:07:08.785199 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.91394 (* 0.0909091 = 0.355813 loss) | |
| I0525 01:07:08.785213 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 4.01526 (* 0.0909091 = 0.365023 loss) | |
| I0525 01:07:08.785228 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.741 (* 0.0909091 = 0.340091 loss) | |
| I0525 01:07:08.785241 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.64358 (* 0.0909091 = 0.331235 loss) | |
| I0525 01:07:08.785254 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.03765 (* 0.0909091 = 0.27615 loss) | |
| I0525 01:07:08.785269 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.14734 (* 0.0909091 = 0.195213 loss) | |
| I0525 01:07:08.785281 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.51079 (* 0.0909091 = 0.137344 loss) | |
| I0525 01:07:08.785295 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.791434 (* 0.0909091 = 0.0719486 loss) | |
| I0525 01:07:08.785310 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.786366 (* 0.0909091 = 0.0714878 loss) | |
| I0525 01:07:08.785322 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.743416 (* 0.0909091 = 0.0675832 loss) | |
| I0525 01:07:08.785336 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.654977 (* 0.0909091 = 0.0595433 loss) | |
| I0525 01:07:08.785351 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.754803 (* 0.0909091 = 0.0686185 loss) | |
| I0525 01:07:08.785363 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.669805 (* 0.0909091 = 0.0608913 loss) | |
| I0525 01:07:08.785377 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.810451 (* 0.0909091 = 0.0736774 loss) | |
| I0525 01:07:08.785392 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0349501 (* 0.0909091 = 0.00317728 loss) | |
| I0525 01:07:08.785405 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0340422 (* 0.0909091 = 0.00309475 loss) | |
| I0525 01:07:08.785419 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00880054 (* 0.0909091 = 0.000800049 loss) | |
| I0525 01:07:08.785434 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00378561 (* 0.0909091 = 0.000344146 loss) | |
| I0525 01:07:08.785447 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00396021 (* 0.0909091 = 0.000360019 loss) | |
| I0525 01:07:08.785461 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00263642 (* 0.0909091 = 0.000239674 loss) | |
| I0525 01:07:08.785475 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00161738 (* 0.0909091 = 0.000147034 loss) | |
| I0525 01:07:08.785490 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.001406 (* 0.0909091 = 0.000127818 loss) | |
| I0525 01:07:08.785502 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:07:08.785514 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:07:08.785526 5272 solver.cpp:245] Train net output #149: total_confidence = 1.35591e-07 | |
| I0525 01:07:08.785548 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 9.73293e-05 | |
| I0525 01:07:08.785562 5272 sgd_solver.cpp:106] Iteration 4500, lr = 0.001 | |
| I0525 01:10:03.713016 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.763 > 30) by scale factor 0.944494 | |
| I0525 01:11:50.603463 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.0353 > 30) by scale factor 0.681272 | |
| I0525 01:13:32.853703 5272 solver.cpp:338] Iteration 5000, Testing net (#0) | |
| I0525 01:14:31.196054 5272 solver.cpp:393] Test loss: 10.8951 | |
| I0525 01:14:31.196226 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0407952 | |
| I0525 01:14:31.196247 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.091 | |
| I0525 01:14:31.196260 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.06 | |
| I0525 01:14:31.196274 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.043 | |
| I0525 01:14:31.196285 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.139 | |
| I0525 01:14:31.196296 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.298 | |
| I0525 01:14:31.196308 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.472 | |
| I0525 01:14:31.196319 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.741 | |
| I0525 01:14:31.196331 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.922 | |
| I0525 01:14:31.196343 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.984 | |
| I0525 01:14:31.196355 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.994 | |
| I0525 01:14:31.196367 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0525 01:14:31.196378 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0525 01:14:31.196389 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0525 01:14:31.196400 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0525 01:14:31.196411 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0525 01:14:31.196422 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0525 01:14:31.196434 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0525 01:14:31.196444 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0525 01:14:31.196455 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0525 01:14:31.196467 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0525 01:14:31.196478 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0525 01:14:31.196490 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0525 01:14:31.196501 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.76091 | |
| I0525 01:14:31.196512 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.162901 | |
| I0525 01:14:31.196528 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 4.18336 (* 0.3 = 1.25501 loss) | |
| I0525 01:14:31.196542 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.08137 (* 0.3 = 0.324412 loss) | |
| I0525 01:14:31.196557 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 3.50257 (* 0.0272727 = 0.0955248 loss) | |
| I0525 01:14:31.196570 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.61815 (* 0.0272727 = 0.0986768 loss) | |
| I0525 01:14:31.196584 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.67538 (* 0.0272727 = 0.100238 loss) | |
| I0525 01:14:31.196599 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.5151 (* 0.0272727 = 0.0958663 loss) | |
| I0525 01:14:31.196611 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 3.03892 (* 0.0272727 = 0.0828796 loss) | |
| I0525 01:14:31.196629 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.51562 (* 0.0272727 = 0.0686077 loss) | |
| I0525 01:14:31.196642 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.48975 (* 0.0272727 = 0.0406296 loss) | |
| I0525 01:14:31.196655 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.581248 (* 0.0272727 = 0.0158522 loss) | |
| I0525 01:14:31.196669 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.170956 (* 0.0272727 = 0.00466244 loss) | |
| I0525 01:14:31.196683 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.103576 (* 0.0272727 = 0.00282481 loss) | |
| I0525 01:14:31.196696 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0561049 (* 0.0272727 = 0.00153013 loss) | |
| I0525 01:14:31.196710 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0401162 (* 0.0272727 = 0.00109408 loss) | |
| I0525 01:14:31.196723 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0430491 (* 0.0272727 = 0.00117407 loss) | |
| I0525 01:14:31.196750 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0316757 (* 0.0272727 = 0.000863884 loss) | |
| I0525 01:14:31.196765 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.026732 (* 0.0272727 = 0.000729054 loss) | |
| I0525 01:14:31.196779 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.020205 (* 0.0272727 = 0.000551045 loss) | |
| I0525 01:14:31.196792 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0142663 (* 0.0272727 = 0.000389082 loss) | |
| I0525 01:14:31.196806 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0140777 (* 0.0272727 = 0.000383936 loss) | |
| I0525 01:14:31.196820 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0128235 (* 0.0272727 = 0.000349732 loss) | |
| I0525 01:14:31.196833 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0143742 (* 0.0272727 = 0.000392025 loss) | |
| I0525 01:14:31.196847 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.014323 (* 0.0272727 = 0.000390627 loss) | |
| I0525 01:14:31.196861 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0134079 (* 0.0272727 = 0.000365669 loss) | |
| I0525 01:14:31.196872 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0392413 | |
| I0525 01:14:31.196884 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.095 | |
| I0525 01:14:31.196897 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.057 | |
| I0525 01:14:31.196907 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.058 | |
| I0525 01:14:31.196918 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.134 | |
| I0525 01:14:31.196933 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.298 | |
| I0525 01:14:31.196944 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.472 | |
| I0525 01:14:31.196956 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.741 | |
| I0525 01:14:31.196967 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.922 | |
| I0525 01:14:31.196979 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.984 | |
| I0525 01:14:31.196990 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.994 | |
| I0525 01:14:31.197001 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0525 01:14:31.197012 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0525 01:14:31.197024 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0525 01:14:31.197036 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0525 01:14:31.197046 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0525 01:14:31.197057 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0525 01:14:31.197068 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0525 01:14:31.197079 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0525 01:14:31.197090 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0525 01:14:31.197101 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0525 01:14:31.197113 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0525 01:14:31.197136 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0525 01:14:31.197149 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.760728 | |
| I0525 01:14:31.197161 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.156201 | |
| I0525 01:14:31.197175 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 4.24936 (* 0.3 = 1.27481 loss) | |
| I0525 01:14:31.197190 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.09876 (* 0.3 = 0.329628 loss) | |
| I0525 01:14:31.197202 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 3.52398 (* 0.0272727 = 0.0961086 loss) | |
| I0525 01:14:31.197216 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.62354 (* 0.0272727 = 0.0988237 loss) | |
| I0525 01:14:31.197229 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.65058 (* 0.0272727 = 0.0995613 loss) | |
| I0525 01:14:31.197254 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 3.51706 (* 0.0272727 = 0.0959199 loss) | |
| I0525 01:14:31.197268 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 3.05321 (* 0.0272727 = 0.0832693 loss) | |
| I0525 01:14:31.197283 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.54011 (* 0.0272727 = 0.0692758 loss) | |
| I0525 01:14:31.197295 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.51205 (* 0.0272727 = 0.0412376 loss) | |
| I0525 01:14:31.197309 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.585978 (* 0.0272727 = 0.0159812 loss) | |
| I0525 01:14:31.197337 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.176286 (* 0.0272727 = 0.00480779 loss) | |
| I0525 01:14:31.197352 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.106274 (* 0.0272727 = 0.00289838 loss) | |
| I0525 01:14:31.197366 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0574618 (* 0.0272727 = 0.00156714 loss) | |
| I0525 01:14:31.197381 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0483056 (* 0.0272727 = 0.00131743 loss) | |
| I0525 01:14:31.197394 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0378245 (* 0.0272727 = 0.00103158 loss) | |
| I0525 01:14:31.197407 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0309816 (* 0.0272727 = 0.000844952 loss) | |
| I0525 01:14:31.197422 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.03057 (* 0.0272727 = 0.000833727 loss) | |
| I0525 01:14:31.197434 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.0270264 (* 0.0272727 = 0.000737084 loss) | |
| I0525 01:14:31.197448 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0206887 (* 0.0272727 = 0.000564238 loss) | |
| I0525 01:14:31.197461 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0206279 (* 0.0272727 = 0.00056258 loss) | |
| I0525 01:14:31.197475 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0137704 (* 0.0272727 = 0.000375556 loss) | |
| I0525 01:14:31.197489 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0170247 (* 0.0272727 = 0.00046431 loss) | |
| I0525 01:14:31.197502 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0130431 (* 0.0272727 = 0.000355722 loss) | |
| I0525 01:14:31.197516 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0137607 (* 0.0272727 = 0.000375293 loss) | |
| I0525 01:14:31.197527 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0611627 | |
| I0525 01:14:31.197540 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.097 | |
| I0525 01:14:31.197551 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.082 | |
| I0525 01:14:31.197561 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.064 | |
| I0525 01:14:31.197573 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.147 | |
| I0525 01:14:31.197584 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.293 | |
| I0525 01:14:31.197594 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.472 | |
| I0525 01:14:31.197607 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.741 | |
| I0525 01:14:31.197618 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.922 | |
| I0525 01:14:31.197629 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.984 | |
| I0525 01:14:31.197640 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.994 | |
| I0525 01:14:31.197651 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0525 01:14:31.197662 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0525 01:14:31.197676 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0525 01:14:31.197687 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0525 01:14:31.197698 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0525 01:14:31.197710 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0525 01:14:31.197721 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0525 01:14:31.197741 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0525 01:14:31.197754 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0525 01:14:31.197765 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0525 01:14:31.197777 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0525 01:14:31.197788 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0525 01:14:31.197798 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.763046 | |
| I0525 01:14:31.197809 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.186017 | |
| I0525 01:14:31.197823 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.53036 (* 1 = 3.53036 loss) | |
| I0525 01:14:31.197839 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.981813 (* 1 = 0.981813 loss) | |
| I0525 01:14:31.197849 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 3.3441 (* 0.0909091 = 0.304009 loss) | |
| I0525 01:14:31.197862 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 3.45409 (* 0.0909091 = 0.314008 loss) | |
| I0525 01:14:31.197876 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.5222 (* 0.0909091 = 0.3202 loss) | |
| I0525 01:14:31.197890 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 3.38196 (* 0.0909091 = 0.307451 loss) | |
| I0525 01:14:31.197902 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.95404 (* 0.0909091 = 0.268549 loss) | |
| I0525 01:14:31.197916 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.45104 (* 0.0909091 = 0.222822 loss) | |
| I0525 01:14:31.197928 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.47183 (* 0.0909091 = 0.133803 loss) | |
| I0525 01:14:31.197942 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.552402 (* 0.0909091 = 0.0502184 loss) | |
| I0525 01:14:31.197954 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.16017 (* 0.0909091 = 0.0145609 loss) | |
| I0525 01:14:31.197968 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0998079 (* 0.0909091 = 0.00907344 loss) | |
| I0525 01:14:31.197984 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0605407 (* 0.0909091 = 0.0055037 loss) | |
| I0525 01:14:31.197999 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0498019 (* 0.0909091 = 0.00452745 loss) | |
| I0525 01:14:31.198012 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0410195 (* 0.0909091 = 0.00372905 loss) | |
| I0525 01:14:31.198025 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.0319506 (* 0.0909091 = 0.0029046 loss) | |
| I0525 01:14:31.198040 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.0252223 (* 0.0909091 = 0.00229293 loss) | |
| I0525 01:14:31.198052 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.017384 (* 0.0909091 = 0.00158036 loss) | |
| I0525 01:14:31.198066 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.009152 (* 0.0909091 = 0.000832 loss) | |
| I0525 01:14:31.198079 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00729522 (* 0.0909091 = 0.000663202 loss) | |
| I0525 01:14:31.198092 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00471872 (* 0.0909091 = 0.000428975 loss) | |
| I0525 01:14:31.198106 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00411859 (* 0.0909091 = 0.000374418 loss) | |
| I0525 01:14:31.198118 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00318579 (* 0.0909091 = 0.000289618 loss) | |
| I0525 01:14:31.198132 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00360116 (* 0.0909091 = 0.000327378 loss) | |
| I0525 01:14:31.198143 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 01:14:31.198154 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:14:31.198164 5272 solver.cpp:406] Test net output #149: total_confidence = 2.65895e-05 | |
| I0525 01:14:31.198176 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000229805 | |
| I0525 01:14:31.198199 5272 solver.cpp:338] Iteration 5000, Testing net (#1) | |
| I0525 01:15:29.690523 5272 solver.cpp:393] Test loss: 11.8057 | |
| I0525 01:15:29.690655 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0328547 | |
| I0525 01:15:29.690682 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.091 | |
| I0525 01:15:29.690708 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.07 | |
| I0525 01:15:29.690724 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.049 | |
| I0525 01:15:29.690737 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.142 | |
| I0525 01:15:29.690749 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.281 | |
| I0525 01:15:29.690762 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.427 | |
| I0525 01:15:29.690773 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.655 | |
| I0525 01:15:29.690784 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.813 | |
| I0525 01:15:29.690796 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.88 | |
| I0525 01:15:29.690807 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.901 | |
| I0525 01:15:29.690819 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.924 | |
| I0525 01:15:29.690830 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.941 | |
| I0525 01:15:29.690843 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.95 | |
| I0525 01:15:29.690855 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.96 | |
| I0525 01:15:29.690865 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.963 | |
| I0525 01:15:29.690881 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.981 | |
| I0525 01:15:29.690892 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.991 | |
| I0525 01:15:29.690904 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.992 | |
| I0525 01:15:29.690917 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.994 | |
| I0525 01:15:29.690927 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.999 | |
| I0525 01:15:29.690938 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.999 | |
| I0525 01:15:29.690949 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.999 | |
| I0525 01:15:29.690963 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.724319 | |
| I0525 01:15:29.690986 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.159007 | |
| I0525 01:15:29.691015 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 4.2718 (* 0.3 = 1.28154 loss) | |
| I0525 01:15:29.691031 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.28014 (* 0.3 = 0.384043 loss) | |
| I0525 01:15:29.691045 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 3.55082 (* 0.0272727 = 0.0968407 loss) | |
| I0525 01:15:29.691059 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.58116 (* 0.0272727 = 0.0976681 loss) | |
| I0525 01:15:29.691072 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.69745 (* 0.0272727 = 0.10084 loss) | |
| I0525 01:15:29.691087 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.51273 (* 0.0272727 = 0.0958017 loss) | |
| I0525 01:15:29.691099 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 3.06272 (* 0.0272727 = 0.0835288 loss) | |
| I0525 01:15:29.691112 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.67913 (* 0.0272727 = 0.0730672 loss) | |
| I0525 01:15:29.691126 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.78855 (* 0.0272727 = 0.0487786 loss) | |
| I0525 01:15:29.691139 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 1.02746 (* 0.0272727 = 0.0280216 loss) | |
| I0525 01:15:29.691153 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.66014 (* 0.0272727 = 0.0180038 loss) | |
| I0525 01:15:29.691165 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.58698 (* 0.0272727 = 0.0160086 loss) | |
| I0525 01:15:29.691179 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.474315 (* 0.0272727 = 0.0129359 loss) | |
| I0525 01:15:29.691192 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.39144 (* 0.0272727 = 0.0106756 loss) | |
| I0525 01:15:29.691206 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.341931 (* 0.0272727 = 0.0093254 loss) | |
| I0525 01:15:29.691241 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.284074 (* 0.0272727 = 0.00774747 loss) | |
| I0525 01:15:29.691256 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.263264 (* 0.0272727 = 0.00717993 loss) | |
| I0525 01:15:29.691268 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.156059 (* 0.0272727 = 0.00425615 loss) | |
| I0525 01:15:29.691282 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.089202 (* 0.0272727 = 0.00243278 loss) | |
| I0525 01:15:29.691296 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0800309 (* 0.0272727 = 0.00218266 loss) | |
| I0525 01:15:29.691309 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0629195 (* 0.0272727 = 0.00171599 loss) | |
| I0525 01:15:29.691323 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0240276 (* 0.0272727 = 0.000655299 loss) | |
| I0525 01:15:29.691336 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0237058 (* 0.0272727 = 0.000646522 loss) | |
| I0525 01:15:29.691352 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0228417 (* 0.0272727 = 0.000622956 loss) | |
| I0525 01:15:29.691375 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0339731 | |
| I0525 01:15:29.691397 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.09 | |
| I0525 01:15:29.691411 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.063 | |
| I0525 01:15:29.691423 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.057 | |
| I0525 01:15:29.691434 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.124 | |
| I0525 01:15:29.691447 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.276 | |
| I0525 01:15:29.691457 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.427 | |
| I0525 01:15:29.691468 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.655 | |
| I0525 01:15:29.691479 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.813 | |
| I0525 01:15:29.691490 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.88 | |
| I0525 01:15:29.691501 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.901 | |
| I0525 01:15:29.691512 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.924 | |
| I0525 01:15:29.691524 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.941 | |
| I0525 01:15:29.691534 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.95 | |
| I0525 01:15:29.691546 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.96 | |
| I0525 01:15:29.691557 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.963 | |
| I0525 01:15:29.691570 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.981 | |
| I0525 01:15:29.691581 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.991 | |
| I0525 01:15:29.691591 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.992 | |
| I0525 01:15:29.691602 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.994 | |
| I0525 01:15:29.691613 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.999 | |
| I0525 01:15:29.691625 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.999 | |
| I0525 01:15:29.691637 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.999 | |
| I0525 01:15:29.691648 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.725273 | |
| I0525 01:15:29.691658 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.158243 | |
| I0525 01:15:29.691675 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 4.35725 (* 0.3 = 1.30718 loss) | |
| I0525 01:15:29.691689 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.31034 (* 0.3 = 0.393101 loss) | |
| I0525 01:15:29.691714 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 3.5784 (* 0.0272727 = 0.0975926 loss) | |
| I0525 01:15:29.691740 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.59583 (* 0.0272727 = 0.0980682 loss) | |
| I0525 01:15:29.691773 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.68726 (* 0.0272727 = 0.100562 loss) | |
| I0525 01:15:29.691787 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 3.52559 (* 0.0272727 = 0.0961524 loss) | |
| I0525 01:15:29.691805 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 3.08803 (* 0.0272727 = 0.084219 loss) | |
| I0525 01:15:29.691831 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.69858 (* 0.0272727 = 0.0735976 loss) | |
| I0525 01:15:29.691854 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.8094 (* 0.0272727 = 0.0493474 loss) | |
| I0525 01:15:29.691869 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 1.02441 (* 0.0272727 = 0.0279383 loss) | |
| I0525 01:15:29.691882 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.677464 (* 0.0272727 = 0.0184763 loss) | |
| I0525 01:15:29.691895 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.584583 (* 0.0272727 = 0.0159432 loss) | |
| I0525 01:15:29.691910 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.486788 (* 0.0272727 = 0.013276 loss) | |
| I0525 01:15:29.691922 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.392599 (* 0.0272727 = 0.0107073 loss) | |
| I0525 01:15:29.691939 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.344252 (* 0.0272727 = 0.00938869 loss) | |
| I0525 01:15:29.691952 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.278492 (* 0.0272727 = 0.00759523 loss) | |
| I0525 01:15:29.691967 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.27383 (* 0.0272727 = 0.0074681 loss) | |
| I0525 01:15:29.691979 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.158455 (* 0.0272727 = 0.0043215 loss) | |
| I0525 01:15:29.691993 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0918658 (* 0.0272727 = 0.00250543 loss) | |
| I0525 01:15:29.692006 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0831915 (* 0.0272727 = 0.00226886 loss) | |
| I0525 01:15:29.692019 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0618085 (* 0.0272727 = 0.00168569 loss) | |
| I0525 01:15:29.692034 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.026766 (* 0.0272727 = 0.000729982 loss) | |
| I0525 01:15:29.692046 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0234775 (* 0.0272727 = 0.000640296 loss) | |
| I0525 01:15:29.692060 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0245529 (* 0.0272727 = 0.000669626 loss) | |
| I0525 01:15:29.692071 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0580329 | |
| I0525 01:15:29.692083 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.094 | |
| I0525 01:15:29.692095 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.077 | |
| I0525 01:15:29.692106 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.084 | |
| I0525 01:15:29.692117 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.139 | |
| I0525 01:15:29.692128 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.279 | |
| I0525 01:15:29.692139 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.427 | |
| I0525 01:15:29.692150 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.655 | |
| I0525 01:15:29.692162 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.813 | |
| I0525 01:15:29.692173 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.88 | |
| I0525 01:15:29.692184 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.901 | |
| I0525 01:15:29.692195 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.924 | |
| I0525 01:15:29.692206 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.941 | |
| I0525 01:15:29.692217 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.95 | |
| I0525 01:15:29.692229 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.96 | |
| I0525 01:15:29.692240 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.963 | |
| I0525 01:15:29.692248 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.981 | |
| I0525 01:15:29.692270 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.991 | |
| I0525 01:15:29.692283 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.992 | |
| I0525 01:15:29.692294 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.994 | |
| I0525 01:15:29.692306 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.999 | |
| I0525 01:15:29.692325 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.999 | |
| I0525 01:15:29.692347 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.999 | |
| I0525 01:15:29.692365 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.729046 | |
| I0525 01:15:29.692378 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.180027 | |
| I0525 01:15:29.692391 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.57124 (* 1 = 3.57124 loss) | |
| I0525 01:15:29.692404 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.12514 (* 1 = 1.12514 loss) | |
| I0525 01:15:29.692419 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 3.39906 (* 0.0909091 = 0.309005 loss) | |
| I0525 01:15:29.692431 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 3.42135 (* 0.0909091 = 0.311032 loss) | |
| I0525 01:15:29.692445 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.54321 (* 0.0909091 = 0.32211 loss) | |
| I0525 01:15:29.692457 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 3.40175 (* 0.0909091 = 0.30925 loss) | |
| I0525 01:15:29.692471 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 3.0095 (* 0.0909091 = 0.273591 loss) | |
| I0525 01:15:29.692483 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.64451 (* 0.0909091 = 0.24041 loss) | |
| I0525 01:15:29.692497 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.7668 (* 0.0909091 = 0.160618 loss) | |
| I0525 01:15:29.692509 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.977754 (* 0.0909091 = 0.0888867 loss) | |
| I0525 01:15:29.692523 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.61344 (* 0.0909091 = 0.0557672 loss) | |
| I0525 01:15:29.692535 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.536173 (* 0.0909091 = 0.048743 loss) | |
| I0525 01:15:29.692548 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.429213 (* 0.0909091 = 0.0390194 loss) | |
| I0525 01:15:29.692562 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.343277 (* 0.0909091 = 0.031207 loss) | |
| I0525 01:15:29.692575 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.312462 (* 0.0909091 = 0.0284056 loss) | |
| I0525 01:15:29.692589 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.261479 (* 0.0909091 = 0.0237708 loss) | |
| I0525 01:15:29.692601 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.256072 (* 0.0909091 = 0.0232793 loss) | |
| I0525 01:15:29.692615 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.142444 (* 0.0909091 = 0.0129495 loss) | |
| I0525 01:15:29.692628 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0785858 (* 0.0909091 = 0.00714416 loss) | |
| I0525 01:15:29.692641 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0754316 (* 0.0909091 = 0.00685742 loss) | |
| I0525 01:15:29.692654 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0613554 (* 0.0909091 = 0.00557776 loss) | |
| I0525 01:15:29.692667 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.014015 (* 0.0909091 = 0.00127409 loss) | |
| I0525 01:15:29.692680 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0138255 (* 0.0909091 = 0.00125686 loss) | |
| I0525 01:15:29.692693 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0138513 (* 0.0909091 = 0.00125921 loss) | |
| I0525 01:15:29.692704 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 01:15:29.692715 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:15:29.692730 5272 solver.cpp:406] Test net output #149: total_confidence = 2.33363e-05 | |
| I0525 01:15:29.692756 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000198021 | |
| I0525 01:15:30.050091 5272 solver.cpp:229] Iteration 5000, loss = 11.7774 | |
| I0525 01:15:30.050176 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0263158 | |
| I0525 01:15:30.050195 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 01:15:30.050209 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 01:15:30.050221 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 01:15:30.050233 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 01:15:30.050246 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0525 01:15:30.050257 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 01:15:30.050271 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 01:15:30.050282 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 01:15:30.050294 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 01:15:30.050307 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 01:15:30.050318 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 01:15:30.050330 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:15:30.050345 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:15:30.050359 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:15:30.050370 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:15:30.050384 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:15:30.050395 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:15:30.050407 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:15:30.050420 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:15:30.050431 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:15:30.050443 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:15:30.050454 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:15:30.050467 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.789773 | |
| I0525 01:15:30.050478 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0789474 | |
| I0525 01:15:30.050495 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.61124 (* 0.3 = 1.08337 loss) | |
| I0525 01:15:30.050509 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.951132 (* 0.3 = 0.28534 loss) | |
| I0525 01:15:30.050523 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.92659 (* 0.0272727 = 0.107089 loss) | |
| I0525 01:15:30.050539 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.55139 (* 0.0272727 = 0.096856 loss) | |
| I0525 01:15:30.050551 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.12024 (* 0.0272727 = 0.11237 loss) | |
| I0525 01:15:30.050565 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.75165 (* 0.0272727 = 0.102318 loss) | |
| I0525 01:15:30.050580 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.09121 (* 0.0272727 = 0.0570331 loss) | |
| I0525 01:15:30.050593 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.0255 (* 0.0272727 = 0.0552408 loss) | |
| I0525 01:15:30.050607 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 0.943919 (* 0.0272727 = 0.0257433 loss) | |
| I0525 01:15:30.050622 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.163657 (* 0.0272727 = 0.00446338 loss) | |
| I0525 01:15:30.050637 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0730641 (* 0.0272727 = 0.00199266 loss) | |
| I0525 01:15:30.050652 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0668919 (* 0.0272727 = 0.00182433 loss) | |
| I0525 01:15:30.050665 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0426135 (* 0.0272727 = 0.00116219 loss) | |
| I0525 01:15:30.050717 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0370983 (* 0.0272727 = 0.00101177 loss) | |
| I0525 01:15:30.050734 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0428163 (* 0.0272727 = 0.00116772 loss) | |
| I0525 01:15:30.050750 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0310513 (* 0.0272727 = 0.000846855 loss) | |
| I0525 01:15:30.050765 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0144841 (* 0.0272727 = 0.000395022 loss) | |
| I0525 01:15:30.050778 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0320794 (* 0.0272727 = 0.000874893 loss) | |
| I0525 01:15:30.050792 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.017461 (* 0.0272727 = 0.000476209 loss) | |
| I0525 01:15:30.050806 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0128127 (* 0.0272727 = 0.000349437 loss) | |
| I0525 01:15:30.050820 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00974174 (* 0.0272727 = 0.000265684 loss) | |
| I0525 01:15:30.050834 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00721823 (* 0.0272727 = 0.000196861 loss) | |
| I0525 01:15:30.050848 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0113255 (* 0.0272727 = 0.000308877 loss) | |
| I0525 01:15:30.050863 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00850658 (* 0.0272727 = 0.000231998 loss) | |
| I0525 01:15:30.050874 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0263158 | |
| I0525 01:15:30.050886 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 01:15:30.050899 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 01:15:30.050910 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:15:30.050922 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 01:15:30.050935 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0525 01:15:30.050946 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 01:15:30.050958 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 01:15:30.050971 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 01:15:30.050982 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 01:15:30.050993 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 01:15:30.051005 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 01:15:30.051017 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:15:30.051028 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:15:30.051040 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:15:30.051051 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:15:30.051062 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:15:30.051074 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:15:30.051085 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:15:30.051096 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:15:30.051108 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:15:30.051120 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:15:30.051131 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:15:30.051142 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.784091 | |
| I0525 01:15:30.051154 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.131579 | |
| I0525 01:15:30.051168 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.80572 (* 0.3 = 1.14172 loss) | |
| I0525 01:15:30.051182 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.983299 (* 0.3 = 0.29499 loss) | |
| I0525 01:15:30.051195 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.68561 (* 0.0272727 = 0.100517 loss) | |
| I0525 01:15:30.051220 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.53513 (* 0.0272727 = 0.0964126 loss) | |
| I0525 01:15:30.051235 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.0635 (* 0.0272727 = 0.110823 loss) | |
| I0525 01:15:30.051249 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.96364 (* 0.0272727 = 0.108099 loss) | |
| I0525 01:15:30.051262 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.15026 (* 0.0272727 = 0.0586434 loss) | |
| I0525 01:15:30.051276 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 1.83167 (* 0.0272727 = 0.0499548 loss) | |
| I0525 01:15:30.051286 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 0.993102 (* 0.0272727 = 0.0270846 loss) | |
| I0525 01:15:30.051296 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.227612 (* 0.0272727 = 0.0062076 loss) | |
| I0525 01:15:30.051312 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.175445 (* 0.0272727 = 0.00478487 loss) | |
| I0525 01:15:30.051326 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0872124 (* 0.0272727 = 0.00237852 loss) | |
| I0525 01:15:30.051340 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0781888 (* 0.0272727 = 0.00213242 loss) | |
| I0525 01:15:30.051354 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.094235 (* 0.0272727 = 0.00257005 loss) | |
| I0525 01:15:30.051367 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0569788 (* 0.0272727 = 0.00155397 loss) | |
| I0525 01:15:30.051381 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0493121 (* 0.0272727 = 0.00134487 loss) | |
| I0525 01:15:30.051399 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0343773 (* 0.0272727 = 0.000937564 loss) | |
| I0525 01:15:30.051414 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0205342 (* 0.0272727 = 0.000560024 loss) | |
| I0525 01:15:30.051429 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0255752 (* 0.0272727 = 0.000697504 loss) | |
| I0525 01:15:30.051442 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0156056 (* 0.0272727 = 0.000425608 loss) | |
| I0525 01:15:30.051455 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0176291 (* 0.0272727 = 0.000480794 loss) | |
| I0525 01:15:30.051470 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0187614 (* 0.0272727 = 0.000511675 loss) | |
| I0525 01:15:30.051483 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00849466 (* 0.0272727 = 0.000231672 loss) | |
| I0525 01:15:30.051497 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0150648 (* 0.0272727 = 0.000410857 loss) | |
| I0525 01:15:30.051509 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0526316 | |
| I0525 01:15:30.051522 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 01:15:30.051533 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:15:30.051544 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:15:30.051556 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 01:15:30.051568 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0525 01:15:30.051580 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 01:15:30.051591 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 01:15:30.051604 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 01:15:30.051615 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 01:15:30.051626 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 01:15:30.051638 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 01:15:30.051650 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:15:30.051661 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:15:30.051672 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:15:30.051695 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:15:30.051708 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:15:30.051719 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:15:30.051731 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:15:30.051743 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:15:30.051754 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:15:30.051766 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:15:30.051777 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:15:30.051789 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.789773 | |
| I0525 01:15:30.051802 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.157895 | |
| I0525 01:15:30.051817 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.6909 (* 1 = 3.6909 loss) | |
| I0525 01:15:30.051831 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.980071 (* 1 = 0.980071 loss) | |
| I0525 01:15:30.051846 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.45222 (* 0.0909091 = 0.313838 loss) | |
| I0525 01:15:30.051859 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.09269 (* 0.0909091 = 0.281154 loss) | |
| I0525 01:15:30.051872 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.75668 (* 0.0909091 = 0.341516 loss) | |
| I0525 01:15:30.051887 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.46372 (* 0.0909091 = 0.314884 loss) | |
| I0525 01:15:30.051900 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.06079 (* 0.0909091 = 0.187345 loss) | |
| I0525 01:15:30.051914 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.88538 (* 0.0909091 = 0.171398 loss) | |
| I0525 01:15:30.051928 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.896321 (* 0.0909091 = 0.0814837 loss) | |
| I0525 01:15:30.051941 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.1649 (* 0.0909091 = 0.0149909 loss) | |
| I0525 01:15:30.051955 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0658796 (* 0.0909091 = 0.00598905 loss) | |
| I0525 01:15:30.051970 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0704603 (* 0.0909091 = 0.00640548 loss) | |
| I0525 01:15:30.051983 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0686437 (* 0.0909091 = 0.00624034 loss) | |
| I0525 01:15:30.051997 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.048662 (* 0.0909091 = 0.00442381 loss) | |
| I0525 01:15:30.052007 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0391164 (* 0.0909091 = 0.00355604 loss) | |
| I0525 01:15:30.052024 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0263761 (* 0.0909091 = 0.00239783 loss) | |
| I0525 01:15:30.052038 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.025218 (* 0.0909091 = 0.00229255 loss) | |
| I0525 01:15:30.052052 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0138686 (* 0.0909091 = 0.00126078 loss) | |
| I0525 01:15:30.052067 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0163264 (* 0.0909091 = 0.00148422 loss) | |
| I0525 01:15:30.052080 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00967876 (* 0.0909091 = 0.000879888 loss) | |
| I0525 01:15:30.052094 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00776347 (* 0.0909091 = 0.00070577 loss) | |
| I0525 01:15:30.052109 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00348213 (* 0.0909091 = 0.000316557 loss) | |
| I0525 01:15:30.052122 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00702622 (* 0.0909091 = 0.000638747 loss) | |
| I0525 01:15:30.052136 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00480392 (* 0.0909091 = 0.00043672 loss) | |
| I0525 01:15:30.052148 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:15:30.052175 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:15:30.052187 5272 solver.cpp:245] Train net output #149: total_confidence = 4.0887e-06 | |
| I0525 01:15:30.052199 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000138115 | |
| I0525 01:15:30.052212 5272 sgd_solver.cpp:106] Iteration 5000, lr = 0.001 | |
| I0525 01:15:55.036623 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.4848 > 30) by scale factor 0.984098 | |
| I0525 01:21:05.942739 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.4828 > 30) by scale factor 0.77957 | |
| I0525 01:21:19.789552 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.1334 > 30) by scale factor 0.695516 | |
| I0525 01:21:54.824642 5272 solver.cpp:229] Iteration 5500, loss = 11.7086 | |
| I0525 01:21:54.824808 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0612245 | |
| I0525 01:21:54.824831 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 01:21:54.824844 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 01:21:54.824856 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 01:21:54.824868 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 01:21:54.824883 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 01:21:54.824897 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 01:21:54.824908 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 01:21:54.824921 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 01:21:54.824934 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 01:21:54.824946 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 01:21:54.824959 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 01:21:54.824970 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:21:54.824982 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:21:54.824995 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:21:54.825006 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:21:54.825018 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:21:54.825031 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:21:54.825042 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:21:54.825053 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:21:54.825065 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:21:54.825076 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:21:54.825088 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:21:54.825099 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0525 01:21:54.825112 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.142857 | |
| I0525 01:21:54.825142 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.8654 (* 0.3 = 1.15962 loss) | |
| I0525 01:21:54.825158 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.20237 (* 0.3 = 0.360711 loss) | |
| I0525 01:21:54.825173 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.17281 (* 0.0272727 = 0.113804 loss) | |
| I0525 01:21:54.825187 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.07229 (* 0.0272727 = 0.111063 loss) | |
| I0525 01:21:54.825201 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.44505 (* 0.0272727 = 0.121229 loss) | |
| I0525 01:21:54.825215 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.8333 (* 0.0272727 = 0.104545 loss) | |
| I0525 01:21:54.825229 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.0852 (* 0.0272727 = 0.0841419 loss) | |
| I0525 01:21:54.825243 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.16553 (* 0.0272727 = 0.0863327 loss) | |
| I0525 01:21:54.825258 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.35628 (* 0.0272727 = 0.0642622 loss) | |
| I0525 01:21:54.825271 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.759412 (* 0.0272727 = 0.0207112 loss) | |
| I0525 01:21:54.825285 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.09411 (* 0.0272727 = 0.0298393 loss) | |
| I0525 01:21:54.825299 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.842985 (* 0.0272727 = 0.0229905 loss) | |
| I0525 01:21:54.825314 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0740819 (* 0.0272727 = 0.00202042 loss) | |
| I0525 01:21:54.825328 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0872207 (* 0.0272727 = 0.00237875 loss) | |
| I0525 01:21:54.825342 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.097248 (* 0.0272727 = 0.00265222 loss) | |
| I0525 01:21:54.825381 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0437392 (* 0.0272727 = 0.00119289 loss) | |
| I0525 01:21:54.825395 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0427714 (* 0.0272727 = 0.00116649 loss) | |
| I0525 01:21:54.825409 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0414018 (* 0.0272727 = 0.00112914 loss) | |
| I0525 01:21:54.825423 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0479304 (* 0.0272727 = 0.00130719 loss) | |
| I0525 01:21:54.825438 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0264771 (* 0.0272727 = 0.000722101 loss) | |
| I0525 01:21:54.825453 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0292656 (* 0.0272727 = 0.000798152 loss) | |
| I0525 01:21:54.825465 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0229532 (* 0.0272727 = 0.000625997 loss) | |
| I0525 01:21:54.825479 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0362936 (* 0.0272727 = 0.000989827 loss) | |
| I0525 01:21:54.825494 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0350493 (* 0.0272727 = 0.000955891 loss) | |
| I0525 01:21:54.825506 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0408163 | |
| I0525 01:21:54.825518 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 01:21:54.825531 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 01:21:54.825542 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:21:54.825553 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 01:21:54.825565 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 01:21:54.825577 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 01:21:54.825589 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 01:21:54.825600 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 01:21:54.825613 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 01:21:54.825624 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 01:21:54.825636 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 01:21:54.825647 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:21:54.825659 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:21:54.825670 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:21:54.825682 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:21:54.825693 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:21:54.825705 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:21:54.825716 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:21:54.825727 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:21:54.825738 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:21:54.825750 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:21:54.825762 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:21:54.825773 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.732955 | |
| I0525 01:21:54.825784 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.142857 | |
| I0525 01:21:54.825798 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 4.28597 (* 0.3 = 1.28579 loss) | |
| I0525 01:21:54.825811 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.40897 (* 0.3 = 0.422692 loss) | |
| I0525 01:21:54.825829 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.92282 (* 0.0272727 = 0.106986 loss) | |
| I0525 01:21:54.825844 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.28764 (* 0.0272727 = 0.116936 loss) | |
| I0525 01:21:54.825868 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.95566 (* 0.0272727 = 0.107882 loss) | |
| I0525 01:21:54.825882 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.81599 (* 0.0272727 = 0.104072 loss) | |
| I0525 01:21:54.825896 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.57889 (* 0.0272727 = 0.097606 loss) | |
| I0525 01:21:54.825911 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.24965 (* 0.0272727 = 0.0886267 loss) | |
| I0525 01:21:54.825927 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.33922 (* 0.0272727 = 0.0637969 loss) | |
| I0525 01:21:54.825942 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.924673 (* 0.0272727 = 0.0252184 loss) | |
| I0525 01:21:54.825955 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.974992 (* 0.0272727 = 0.0265907 loss) | |
| I0525 01:21:54.825966 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.777068 (* 0.0272727 = 0.0211928 loss) | |
| I0525 01:21:54.825976 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0911114 (* 0.0272727 = 0.00248486 loss) | |
| I0525 01:21:54.825990 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0619259 (* 0.0272727 = 0.00168889 loss) | |
| I0525 01:21:54.826004 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.051186 (* 0.0272727 = 0.00139598 loss) | |
| I0525 01:21:54.826017 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0393038 (* 0.0272727 = 0.00107192 loss) | |
| I0525 01:21:54.826031 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0370045 (* 0.0272727 = 0.00100921 loss) | |
| I0525 01:21:54.826045 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0222412 (* 0.0272727 = 0.000606579 loss) | |
| I0525 01:21:54.826059 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0186889 (* 0.0272727 = 0.000509697 loss) | |
| I0525 01:21:54.826073 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0208906 (* 0.0272727 = 0.000569745 loss) | |
| I0525 01:21:54.826086 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0145325 (* 0.0272727 = 0.000396342 loss) | |
| I0525 01:21:54.826100 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0134465 (* 0.0272727 = 0.000366724 loss) | |
| I0525 01:21:54.826114 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0175038 (* 0.0272727 = 0.000477378 loss) | |
| I0525 01:21:54.826128 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.014018 (* 0.0272727 = 0.000382308 loss) | |
| I0525 01:21:54.826140 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0816327 | |
| I0525 01:21:54.826153 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 01:21:54.826164 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:21:54.826176 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:21:54.826189 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 01:21:54.826200 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 01:21:54.826211 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 01:21:54.826223 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 01:21:54.826236 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 01:21:54.826247 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 01:21:54.826258 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 01:21:54.826270 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 01:21:54.826282 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:21:54.826293 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:21:54.826304 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:21:54.826315 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:21:54.826328 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:21:54.826349 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:21:54.826361 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:21:54.826373 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:21:54.826385 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:21:54.826396 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:21:54.826407 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:21:54.826419 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.715909 | |
| I0525 01:21:54.826431 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.204082 | |
| I0525 01:21:54.826444 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 4.38161 (* 1 = 4.38161 loss) | |
| I0525 01:21:54.826458 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.6397 (* 1 = 1.6397 loss) | |
| I0525 01:21:54.826472 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.85716 (* 0.0909091 = 0.350651 loss) | |
| I0525 01:21:54.826485 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 4.16593 (* 0.0909091 = 0.378721 loss) | |
| I0525 01:21:54.826499 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.03415 (* 0.0909091 = 0.366741 loss) | |
| I0525 01:21:54.826514 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.70777 (* 0.0909091 = 0.33707 loss) | |
| I0525 01:21:54.826526 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.58276 (* 0.0909091 = 0.325705 loss) | |
| I0525 01:21:54.826540 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.42536 (* 0.0909091 = 0.311396 loss) | |
| I0525 01:21:54.826553 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.17362 (* 0.0909091 = 0.197602 loss) | |
| I0525 01:21:54.826567 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.801666 (* 0.0909091 = 0.0728788 loss) | |
| I0525 01:21:54.826581 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 1.27326 (* 0.0909091 = 0.115751 loss) | |
| I0525 01:21:54.826594 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.933234 (* 0.0909091 = 0.0848395 loss) | |
| I0525 01:21:54.826608 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.193865 (* 0.0909091 = 0.0176241 loss) | |
| I0525 01:21:54.826622 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.185962 (* 0.0909091 = 0.0169057 loss) | |
| I0525 01:21:54.826635 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.166974 (* 0.0909091 = 0.0151795 loss) | |
| I0525 01:21:54.826649 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0826608 (* 0.0909091 = 0.00751462 loss) | |
| I0525 01:21:54.826663 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0643418 (* 0.0909091 = 0.00584926 loss) | |
| I0525 01:21:54.826676 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0505717 (* 0.0909091 = 0.00459743 loss) | |
| I0525 01:21:54.826690 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0123263 (* 0.0909091 = 0.00112057 loss) | |
| I0525 01:21:54.826704 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00664821 (* 0.0909091 = 0.000604382 loss) | |
| I0525 01:21:54.826719 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00309728 (* 0.0909091 = 0.000281571 loss) | |
| I0525 01:21:54.826731 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00375836 (* 0.0909091 = 0.000341669 loss) | |
| I0525 01:21:54.826745 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00203365 (* 0.0909091 = 0.000184877 loss) | |
| I0525 01:21:54.826762 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00190314 (* 0.0909091 = 0.000173013 loss) | |
| I0525 01:21:54.826776 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:21:54.826786 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:21:54.826797 5272 solver.cpp:245] Train net output #149: total_confidence = 1.11675e-07 | |
| I0525 01:21:54.826819 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 4.57212e-05 | |
| I0525 01:21:54.826833 5272 sgd_solver.cpp:106] Iteration 5500, lr = 0.001 | |
| I0525 01:23:22.183063 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.2565 > 30) by scale factor 0.875746 | |
| I0525 01:24:37.573218 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.1512 > 30) by scale factor 0.963044 | |
| I0525 01:26:52.223706 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.1043 > 30) by scale factor 0.879654 | |
| I0525 01:28:19.594398 5272 solver.cpp:229] Iteration 6000, loss = 11.6458 | |
| I0525 01:28:19.594529 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0425532 | |
| I0525 01:28:19.594549 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 01:28:19.594563 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 01:28:19.594574 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 01:28:19.594586 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 01:28:19.594599 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 01:28:19.594611 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 01:28:19.594624 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 01:28:19.594635 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 01:28:19.594647 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 01:28:19.594660 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 01:28:19.594672 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 01:28:19.594684 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:28:19.594696 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:28:19.594707 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:28:19.594719 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:28:19.594732 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:28:19.594743 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:28:19.594754 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:28:19.594766 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:28:19.594779 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:28:19.594790 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:28:19.594802 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:28:19.594815 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0525 01:28:19.594826 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.148936 | |
| I0525 01:28:19.594842 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.6204 (* 0.3 = 1.08612 loss) | |
| I0525 01:28:19.594856 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.14304 (* 0.3 = 0.342911 loss) | |
| I0525 01:28:19.594871 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.12695 (* 0.0272727 = 0.112553 loss) | |
| I0525 01:28:19.594887 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.25012 (* 0.0272727 = 0.0886398 loss) | |
| I0525 01:28:19.594902 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.52438 (* 0.0272727 = 0.0961196 loss) | |
| I0525 01:28:19.594915 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.51415 (* 0.0272727 = 0.0958404 loss) | |
| I0525 01:28:19.594929 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.05499 (* 0.0272727 = 0.083318 loss) | |
| I0525 01:28:19.594944 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.43731 (* 0.0272727 = 0.0664722 loss) | |
| I0525 01:28:19.594956 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.44943 (* 0.0272727 = 0.0668025 loss) | |
| I0525 01:28:19.594970 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.614644 (* 0.0272727 = 0.016763 loss) | |
| I0525 01:28:19.594985 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.532718 (* 0.0272727 = 0.0145287 loss) | |
| I0525 01:28:19.595000 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0448328 (* 0.0272727 = 0.00122271 loss) | |
| I0525 01:28:19.595013 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0357748 (* 0.0272727 = 0.000975677 loss) | |
| I0525 01:28:19.595027 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0491628 (* 0.0272727 = 0.0013408 loss) | |
| I0525 01:28:19.595041 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0249795 (* 0.0272727 = 0.00068126 loss) | |
| I0525 01:28:19.595075 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0133067 (* 0.0272727 = 0.000362909 loss) | |
| I0525 01:28:19.595091 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0192937 (* 0.0272727 = 0.000526192 loss) | |
| I0525 01:28:19.595105 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0093134 (* 0.0272727 = 0.000254002 loss) | |
| I0525 01:28:19.595119 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0102828 (* 0.0272727 = 0.00028044 loss) | |
| I0525 01:28:19.595132 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0245455 (* 0.0272727 = 0.000669424 loss) | |
| I0525 01:28:19.595146 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0282189 (* 0.0272727 = 0.000769606 loss) | |
| I0525 01:28:19.595160 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0182786 (* 0.0272727 = 0.000498507 loss) | |
| I0525 01:28:19.595175 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0268492 (* 0.0272727 = 0.000732251 loss) | |
| I0525 01:28:19.595188 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0229486 (* 0.0272727 = 0.000625871 loss) | |
| I0525 01:28:19.595201 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 01:28:19.595213 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 01:28:19.595226 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 01:28:19.595237 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:28:19.595248 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0525 01:28:19.595260 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 01:28:19.595273 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 01:28:19.595283 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 01:28:19.595295 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 01:28:19.595307 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 01:28:19.595319 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 01:28:19.595330 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 01:28:19.595342 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:28:19.595353 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:28:19.595366 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:28:19.595376 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:28:19.595387 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:28:19.595399 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:28:19.595410 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:28:19.595422 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:28:19.595433 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:28:19.595445 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:28:19.595456 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:28:19.595468 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0525 01:28:19.595479 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.12766 | |
| I0525 01:28:19.595494 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.54078 (* 0.3 = 1.06224 loss) | |
| I0525 01:28:19.595507 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.15198 (* 0.3 = 0.345593 loss) | |
| I0525 01:28:19.595521 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.96571 (* 0.0272727 = 0.108156 loss) | |
| I0525 01:28:19.595535 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.00419 (* 0.0272727 = 0.109205 loss) | |
| I0525 01:28:19.595563 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.92938 (* 0.0272727 = 0.107165 loss) | |
| I0525 01:28:19.595578 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.21865 (* 0.0272727 = 0.0877814 loss) | |
| I0525 01:28:19.595592 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.18609 (* 0.0272727 = 0.0868933 loss) | |
| I0525 01:28:19.595607 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.1694 (* 0.0272727 = 0.0591656 loss) | |
| I0525 01:28:19.595620 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.70447 (* 0.0272727 = 0.0737584 loss) | |
| I0525 01:28:19.595633 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.872938 (* 0.0272727 = 0.0238074 loss) | |
| I0525 01:28:19.595648 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.760215 (* 0.0272727 = 0.0207331 loss) | |
| I0525 01:28:19.595661 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0151378 (* 0.0272727 = 0.00041285 loss) | |
| I0525 01:28:19.595675 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.014434 (* 0.0272727 = 0.000393655 loss) | |
| I0525 01:28:19.595690 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0217488 (* 0.0272727 = 0.000593148 loss) | |
| I0525 01:28:19.595703 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00969083 (* 0.0272727 = 0.000264295 loss) | |
| I0525 01:28:19.595717 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0101362 (* 0.0272727 = 0.000276442 loss) | |
| I0525 01:28:19.595731 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00463198 (* 0.0272727 = 0.000126327 loss) | |
| I0525 01:28:19.595744 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00762395 (* 0.0272727 = 0.000207926 loss) | |
| I0525 01:28:19.595758 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00534835 (* 0.0272727 = 0.000145864 loss) | |
| I0525 01:28:19.595772 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00569893 (* 0.0272727 = 0.000155425 loss) | |
| I0525 01:28:19.595787 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00596906 (* 0.0272727 = 0.000162792 loss) | |
| I0525 01:28:19.595800 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00305128 (* 0.0272727 = 8.32166e-05 loss) | |
| I0525 01:28:19.595810 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00458369 (* 0.0272727 = 0.00012501 loss) | |
| I0525 01:28:19.595820 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00339497 (* 0.0272727 = 9.259e-05 loss) | |
| I0525 01:28:19.595834 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0212766 | |
| I0525 01:28:19.595845 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 01:28:19.595857 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:28:19.595870 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 01:28:19.595881 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 01:28:19.595892 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 01:28:19.595904 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 01:28:19.595916 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 01:28:19.595930 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 01:28:19.595942 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 01:28:19.595954 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 01:28:19.595966 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 01:28:19.595978 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:28:19.595988 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:28:19.596000 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:28:19.596011 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:28:19.596024 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:28:19.596045 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:28:19.596057 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:28:19.596070 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:28:19.596081 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:28:19.596091 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:28:19.596103 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:28:19.596114 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0525 01:28:19.596127 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.12766 | |
| I0525 01:28:19.596140 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.4685 (* 1 = 3.4685 loss) | |
| I0525 01:28:19.596153 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.01921 (* 1 = 1.01921 loss) | |
| I0525 01:28:19.596168 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.54116 (* 0.0909091 = 0.321923 loss) | |
| I0525 01:28:19.596180 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.04179 (* 0.0909091 = 0.276527 loss) | |
| I0525 01:28:19.596194 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.89629 (* 0.0909091 = 0.263299 loss) | |
| I0525 01:28:19.596207 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.19986 (* 0.0909091 = 0.290896 loss) | |
| I0525 01:28:19.596220 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.76122 (* 0.0909091 = 0.25102 loss) | |
| I0525 01:28:19.596235 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.09655 (* 0.0909091 = 0.190595 loss) | |
| I0525 01:28:19.596247 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.47341 (* 0.0909091 = 0.224855 loss) | |
| I0525 01:28:19.596261 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.688439 (* 0.0909091 = 0.0625854 loss) | |
| I0525 01:28:19.596274 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.506619 (* 0.0909091 = 0.0460563 loss) | |
| I0525 01:28:19.596288 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0386434 (* 0.0909091 = 0.00351304 loss) | |
| I0525 01:28:19.596302 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0174129 (* 0.0909091 = 0.00158299 loss) | |
| I0525 01:28:19.596315 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.027014 (* 0.0909091 = 0.00245581 loss) | |
| I0525 01:28:19.596329 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0181237 (* 0.0909091 = 0.00164761 loss) | |
| I0525 01:28:19.596343 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0156944 (* 0.0909091 = 0.00142676 loss) | |
| I0525 01:28:19.596356 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0159921 (* 0.0909091 = 0.00145383 loss) | |
| I0525 01:28:19.596370 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00795425 (* 0.0909091 = 0.000723114 loss) | |
| I0525 01:28:19.596384 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00478161 (* 0.0909091 = 0.000434691 loss) | |
| I0525 01:28:19.596397 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00284472 (* 0.0909091 = 0.000258611 loss) | |
| I0525 01:28:19.596410 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00146289 (* 0.0909091 = 0.00013299 loss) | |
| I0525 01:28:19.596424 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00146997 (* 0.0909091 = 0.000133633 loss) | |
| I0525 01:28:19.596438 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00141295 (* 0.0909091 = 0.00012845 loss) | |
| I0525 01:28:19.596452 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000765025 (* 0.0909091 = 6.95478e-05 loss) | |
| I0525 01:28:19.596464 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:28:19.596477 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:28:19.596487 5272 solver.cpp:245] Train net output #149: total_confidence = 6.92402e-05 | |
| I0525 01:28:19.596504 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000671815 | |
| I0525 01:28:19.596520 5272 sgd_solver.cpp:106] Iteration 6000, lr = 0.001 | |
| I0525 01:30:43.841531 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 52.7029 > 30) by scale factor 0.569229 | |
| I0525 01:31:18.461650 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.6182 > 30) by scale factor 0.842266 | |
| I0525 01:32:09.255059 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.1813 > 30) by scale factor 0.993994 | |
| I0525 01:32:56.974020 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.8621 > 30) by scale factor 0.912906 | |
| I0525 01:34:44.327571 5272 solver.cpp:229] Iteration 6500, loss = 11.2361 | |
| I0525 01:34:44.327715 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.133333 | |
| I0525 01:34:44.327736 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 01:34:44.327749 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 01:34:44.327762 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 01:34:44.327775 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 01:34:44.327786 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 01:34:44.327800 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 01:34:44.327811 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 01:34:44.327823 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 01:34:44.327836 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 01:34:44.327847 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 01:34:44.327859 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 01:34:44.327872 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:34:44.327886 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:34:44.327898 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:34:44.327911 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:34:44.327924 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:34:44.327935 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:34:44.327947 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:34:44.327958 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:34:44.327970 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:34:44.327982 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:34:44.327993 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:34:44.328006 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.755682 | |
| I0525 01:34:44.328017 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.266667 | |
| I0525 01:34:44.328032 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.45076 (* 0.3 = 1.03523 loss) | |
| I0525 01:34:44.328047 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.15443 (* 0.3 = 0.34633 loss) | |
| I0525 01:34:44.328063 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.3358 (* 0.0272727 = 0.0909764 loss) | |
| I0525 01:34:44.328076 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.78348 (* 0.0272727 = 0.103186 loss) | |
| I0525 01:34:44.328090 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.83739 (* 0.0272727 = 0.104656 loss) | |
| I0525 01:34:44.328104 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.13074 (* 0.0272727 = 0.0853839 loss) | |
| I0525 01:34:44.328117 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.43036 (* 0.0272727 = 0.0935552 loss) | |
| I0525 01:34:44.328132 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.89787 (* 0.0272727 = 0.0790328 loss) | |
| I0525 01:34:44.328146 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 0.792108 (* 0.0272727 = 0.021603 loss) | |
| I0525 01:34:44.328161 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.817986 (* 0.0272727 = 0.0223087 loss) | |
| I0525 01:34:44.328176 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.173852 (* 0.0272727 = 0.00474141 loss) | |
| I0525 01:34:44.328189 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.103415 (* 0.0272727 = 0.0028204 loss) | |
| I0525 01:34:44.328204 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.112341 (* 0.0272727 = 0.00306384 loss) | |
| I0525 01:34:44.328218 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.036446 (* 0.0272727 = 0.000993982 loss) | |
| I0525 01:34:44.328233 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0270094 (* 0.0272727 = 0.000736619 loss) | |
| I0525 01:34:44.328270 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0250317 (* 0.0272727 = 0.000682681 loss) | |
| I0525 01:34:44.328285 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0127098 (* 0.0272727 = 0.00034663 loss) | |
| I0525 01:34:44.328299 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0124264 (* 0.0272727 = 0.000338901 loss) | |
| I0525 01:34:44.328313 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0096954 (* 0.0272727 = 0.00026442 loss) | |
| I0525 01:34:44.328328 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00476268 (* 0.0272727 = 0.000129891 loss) | |
| I0525 01:34:44.328342 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00552775 (* 0.0272727 = 0.000150757 loss) | |
| I0525 01:34:44.328356 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00390874 (* 0.0272727 = 0.000106602 loss) | |
| I0525 01:34:44.328371 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00294472 (* 0.0272727 = 8.03107e-05 loss) | |
| I0525 01:34:44.328384 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00363078 (* 0.0272727 = 9.90213e-05 loss) | |
| I0525 01:34:44.328397 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.155556 | |
| I0525 01:34:44.328409 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 01:34:44.328420 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 01:34:44.328433 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:34:44.328444 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0525 01:34:44.328456 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 01:34:44.328469 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 01:34:44.328480 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 01:34:44.328492 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 01:34:44.328503 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 01:34:44.328516 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 01:34:44.328526 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 01:34:44.328538 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:34:44.328549 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:34:44.328560 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:34:44.328572 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:34:44.328583 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:34:44.328595 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:34:44.328606 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:34:44.328618 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:34:44.328629 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:34:44.328641 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:34:44.328652 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:34:44.328665 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.772727 | |
| I0525 01:34:44.328675 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.266667 | |
| I0525 01:34:44.328690 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.41064 (* 0.3 = 1.02319 loss) | |
| I0525 01:34:44.328703 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.07871 (* 0.3 = 0.323612 loss) | |
| I0525 01:34:44.328717 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.72189 (* 0.0272727 = 0.101506 loss) | |
| I0525 01:34:44.328734 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.84512 (* 0.0272727 = 0.104867 loss) | |
| I0525 01:34:44.328759 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.54188 (* 0.0272727 = 0.0965967 loss) | |
| I0525 01:34:44.328774 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.23111 (* 0.0272727 = 0.0881211 loss) | |
| I0525 01:34:44.328789 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.72981 (* 0.0272727 = 0.101722 loss) | |
| I0525 01:34:44.328802 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.21684 (* 0.0272727 = 0.0877319 loss) | |
| I0525 01:34:44.328816 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 0.959989 (* 0.0272727 = 0.0261815 loss) | |
| I0525 01:34:44.328830 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.04457 (* 0.0272727 = 0.0284882 loss) | |
| I0525 01:34:44.328845 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.192759 (* 0.0272727 = 0.00525706 loss) | |
| I0525 01:34:44.328858 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.178468 (* 0.0272727 = 0.0048673 loss) | |
| I0525 01:34:44.328872 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0993469 (* 0.0272727 = 0.00270946 loss) | |
| I0525 01:34:44.328886 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0654412 (* 0.0272727 = 0.00178476 loss) | |
| I0525 01:34:44.328902 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.150113 (* 0.0272727 = 0.004094 loss) | |
| I0525 01:34:44.328915 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.060924 (* 0.0272727 = 0.00166156 loss) | |
| I0525 01:34:44.328932 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0358927 (* 0.0272727 = 0.000978892 loss) | |
| I0525 01:34:44.328946 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0247097 (* 0.0272727 = 0.000673901 loss) | |
| I0525 01:34:44.328960 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0160443 (* 0.0272727 = 0.000437573 loss) | |
| I0525 01:34:44.328974 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00849203 (* 0.0272727 = 0.000231601 loss) | |
| I0525 01:34:44.328987 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00995381 (* 0.0272727 = 0.000271468 loss) | |
| I0525 01:34:44.329001 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00575897 (* 0.0272727 = 0.000157063 loss) | |
| I0525 01:34:44.329015 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00569467 (* 0.0272727 = 0.000155309 loss) | |
| I0525 01:34:44.329028 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0170782 (* 0.0272727 = 0.00046577 loss) | |
| I0525 01:34:44.329041 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0888889 | |
| I0525 01:34:44.329053 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 01:34:44.329064 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 01:34:44.329077 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:34:44.329087 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 01:34:44.329099 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 01:34:44.329112 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 01:34:44.329138 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 01:34:44.329151 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 01:34:44.329164 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 01:34:44.329175 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 01:34:44.329187 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 01:34:44.329200 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:34:44.329210 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:34:44.329222 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:34:44.329233 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:34:44.329246 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:34:44.329267 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:34:44.329282 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:34:44.329289 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:34:44.329298 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:34:44.329309 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:34:44.329320 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:34:44.329332 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0525 01:34:44.329344 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.222222 | |
| I0525 01:34:44.329357 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.35219 (* 1 = 3.35219 loss) | |
| I0525 01:34:44.329371 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.988062 (* 1 = 0.988062 loss) | |
| I0525 01:34:44.329385 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.10535 (* 0.0909091 = 0.282305 loss) | |
| I0525 01:34:44.329398 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.40507 (* 0.0909091 = 0.309552 loss) | |
| I0525 01:34:44.329412 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.54959 (* 0.0909091 = 0.32269 loss) | |
| I0525 01:34:44.329426 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.03632 (* 0.0909091 = 0.276029 loss) | |
| I0525 01:34:44.329439 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.35887 (* 0.0909091 = 0.305352 loss) | |
| I0525 01:34:44.329452 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.60731 (* 0.0909091 = 0.237028 loss) | |
| I0525 01:34:44.329466 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.718144 (* 0.0909091 = 0.0652858 loss) | |
| I0525 01:34:44.329480 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.819801 (* 0.0909091 = 0.0745274 loss) | |
| I0525 01:34:44.329494 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0457272 (* 0.0909091 = 0.00415702 loss) | |
| I0525 01:34:44.329509 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0265212 (* 0.0909091 = 0.00241102 loss) | |
| I0525 01:34:44.329522 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0336291 (* 0.0909091 = 0.0030572 loss) | |
| I0525 01:34:44.329536 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0267228 (* 0.0909091 = 0.00242934 loss) | |
| I0525 01:34:44.329550 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0171956 (* 0.0909091 = 0.00156324 loss) | |
| I0525 01:34:44.329565 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0111593 (* 0.0909091 = 0.00101448 loss) | |
| I0525 01:34:44.329577 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0125119 (* 0.0909091 = 0.00113745 loss) | |
| I0525 01:34:44.329591 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00551288 (* 0.0909091 = 0.000501171 loss) | |
| I0525 01:34:44.329605 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0039477 (* 0.0909091 = 0.000358882 loss) | |
| I0525 01:34:44.329619 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00226272 (* 0.0909091 = 0.000205702 loss) | |
| I0525 01:34:44.329633 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00140612 (* 0.0909091 = 0.000127829 loss) | |
| I0525 01:34:44.329646 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00132471 (* 0.0909091 = 0.000120428 loss) | |
| I0525 01:34:44.329660 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000658678 (* 0.0909091 = 5.98798e-05 loss) | |
| I0525 01:34:44.329674 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000450713 (* 0.0909091 = 4.09739e-05 loss) | |
| I0525 01:34:44.329686 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:34:44.329697 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:34:44.329708 5272 solver.cpp:245] Train net output #149: total_confidence = 7.13444e-06 | |
| I0525 01:34:44.329730 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000192928 | |
| I0525 01:34:44.329744 5272 sgd_solver.cpp:106] Iteration 6500, lr = 0.001 | |
| I0525 01:35:16.247216 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.6775 > 30) by scale factor 0.796232 | |
| I0525 01:35:32.398084 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.5412 > 30) by scale factor 0.951138 | |
| I0525 01:36:41.681705 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.9593 > 30) by scale factor 0.938694 | |
| I0525 01:37:53.278947 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.8261 > 30) by scale factor 0.684524 | |
| I0525 01:39:02.571017 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.9009 > 30) by scale factor 0.970847 | |
| I0525 01:40:32.632555 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.3401 > 30) by scale factor 0.988789 | |
| I0525 01:41:09.211504 5272 solver.cpp:229] Iteration 7000, loss = 11.1777 | |
| I0525 01:41:09.211596 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0196078 | |
| I0525 01:41:09.211614 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 01:41:09.211627 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 01:41:09.211639 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0525 01:41:09.211652 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0525 01:41:09.211663 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 01:41:09.211675 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 01:41:09.211688 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0525 01:41:09.211699 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0525 01:41:09.211710 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 01:41:09.211722 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 01:41:09.211733 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 01:41:09.211745 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:41:09.211756 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:41:09.211767 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:41:09.211781 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:41:09.211805 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:41:09.211829 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:41:09.211843 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:41:09.211853 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:41:09.211865 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:41:09.211877 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:41:09.211889 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:41:09.211900 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0525 01:41:09.211912 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.196078 | |
| I0525 01:41:09.211927 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.39343 (* 0.3 = 1.01803 loss) | |
| I0525 01:41:09.211942 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.10281 (* 0.3 = 0.330842 loss) | |
| I0525 01:41:09.211956 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.37448 (* 0.0272727 = 0.0920312 loss) | |
| I0525 01:41:09.211969 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.38246 (* 0.0272727 = 0.092249 loss) | |
| I0525 01:41:09.211983 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.71065 (* 0.0272727 = 0.101199 loss) | |
| I0525 01:41:09.211997 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.82135 (* 0.0272727 = 0.076946 loss) | |
| I0525 01:41:09.212010 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.98644 (* 0.0272727 = 0.0814485 loss) | |
| I0525 01:41:09.212024 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.09287 (* 0.0272727 = 0.084351 loss) | |
| I0525 01:41:09.212038 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 3.53398 (* 0.0272727 = 0.0963812 loss) | |
| I0525 01:41:09.212051 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 2.37116 (* 0.0272727 = 0.0646681 loss) | |
| I0525 01:41:09.212065 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0150004 (* 0.0272727 = 0.000409103 loss) | |
| I0525 01:41:09.212080 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0197339 (* 0.0272727 = 0.000538198 loss) | |
| I0525 01:41:09.212093 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0107874 (* 0.0272727 = 0.000294201 loss) | |
| I0525 01:41:09.212107 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00916167 (* 0.0272727 = 0.000249864 loss) | |
| I0525 01:41:09.212124 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00963324 (* 0.0272727 = 0.000262725 loss) | |
| I0525 01:41:09.212157 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00448462 (* 0.0272727 = 0.000122308 loss) | |
| I0525 01:41:09.212172 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00602586 (* 0.0272727 = 0.000164342 loss) | |
| I0525 01:41:09.212187 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00489558 (* 0.0272727 = 0.000133516 loss) | |
| I0525 01:41:09.212200 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00264398 (* 0.0272727 = 7.21087e-05 loss) | |
| I0525 01:41:09.212213 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00129137 (* 0.0272727 = 3.52192e-05 loss) | |
| I0525 01:41:09.212227 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00271744 (* 0.0272727 = 7.41119e-05 loss) | |
| I0525 01:41:09.212241 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00293754 (* 0.0272727 = 8.01148e-05 loss) | |
| I0525 01:41:09.212255 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00190446 (* 0.0272727 = 5.19399e-05 loss) | |
| I0525 01:41:09.212268 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000886508 (* 0.0272727 = 2.41775e-05 loss) | |
| I0525 01:41:09.212280 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0 | |
| I0525 01:41:09.212292 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 01:41:09.212303 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 01:41:09.212314 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:41:09.212327 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 01:41:09.212337 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 01:41:09.212349 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 01:41:09.212362 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0525 01:41:09.212373 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0525 01:41:09.212384 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 01:41:09.212395 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 01:41:09.212407 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 01:41:09.212419 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:41:09.212430 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:41:09.212440 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:41:09.212451 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:41:09.212462 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:41:09.212474 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:41:09.212486 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:41:09.212496 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:41:09.212507 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:41:09.212518 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:41:09.212529 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:41:09.212540 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0525 01:41:09.212553 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.156863 | |
| I0525 01:41:09.212565 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.47728 (* 0.3 = 1.04318 loss) | |
| I0525 01:41:09.212579 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.09801 (* 0.3 = 0.329402 loss) | |
| I0525 01:41:09.212592 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.32217 (* 0.0272727 = 0.0906046 loss) | |
| I0525 01:41:09.212606 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.69972 (* 0.0272727 = 0.100901 loss) | |
| I0525 01:41:09.212630 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.39013 (* 0.0272727 = 0.0924582 loss) | |
| I0525 01:41:09.212644 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.27105 (* 0.0272727 = 0.0892104 loss) | |
| I0525 01:41:09.212658 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.13515 (* 0.0272727 = 0.0855041 loss) | |
| I0525 01:41:09.212671 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.51957 (* 0.0272727 = 0.0959884 loss) | |
| I0525 01:41:09.212684 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 3.30441 (* 0.0272727 = 0.0901203 loss) | |
| I0525 01:41:09.212698 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 2.91285 (* 0.0272727 = 0.0794414 loss) | |
| I0525 01:41:09.212712 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0204903 (* 0.0272727 = 0.000558826 loss) | |
| I0525 01:41:09.212725 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0145741 (* 0.0272727 = 0.000397476 loss) | |
| I0525 01:41:09.212739 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0110597 (* 0.0272727 = 0.000301628 loss) | |
| I0525 01:41:09.212752 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0118365 (* 0.0272727 = 0.000322814 loss) | |
| I0525 01:41:09.212766 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00944735 (* 0.0272727 = 0.000257655 loss) | |
| I0525 01:41:09.212779 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00581357 (* 0.0272727 = 0.000158552 loss) | |
| I0525 01:41:09.212793 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00653237 (* 0.0272727 = 0.000178155 loss) | |
| I0525 01:41:09.212807 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00480134 (* 0.0272727 = 0.000130946 loss) | |
| I0525 01:41:09.212821 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00313386 (* 0.0272727 = 8.54688e-05 loss) | |
| I0525 01:41:09.212834 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00282644 (* 0.0272727 = 7.70848e-05 loss) | |
| I0525 01:41:09.212848 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00308487 (* 0.0272727 = 8.41327e-05 loss) | |
| I0525 01:41:09.212864 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00252112 (* 0.0272727 = 6.87578e-05 loss) | |
| I0525 01:41:09.212879 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00204601 (* 0.0272727 = 5.58003e-05 loss) | |
| I0525 01:41:09.212893 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00272406 (* 0.0272727 = 7.42925e-05 loss) | |
| I0525 01:41:09.212905 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0784314 | |
| I0525 01:41:09.212918 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 01:41:09.212929 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:41:09.212939 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:41:09.212951 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 01:41:09.212962 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 01:41:09.212975 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 01:41:09.212985 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0525 01:41:09.212997 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0525 01:41:09.213008 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 01:41:09.213021 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 01:41:09.213032 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 01:41:09.213042 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:41:09.213054 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:41:09.213065 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:41:09.213076 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:41:09.213088 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:41:09.213107 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:41:09.213135 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:41:09.213150 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:41:09.213161 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:41:09.213172 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:41:09.213183 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:41:09.213194 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.721591 | |
| I0525 01:41:09.213207 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.294118 | |
| I0525 01:41:09.213220 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.05322 (* 1 = 3.05322 loss) | |
| I0525 01:41:09.213233 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.966724 (* 1 = 0.966724 loss) | |
| I0525 01:41:09.213248 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.98365 (* 0.0909091 = 0.271241 loss) | |
| I0525 01:41:09.213261 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.30543 (* 0.0909091 = 0.300493 loss) | |
| I0525 01:41:09.213274 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.15522 (* 0.0909091 = 0.286839 loss) | |
| I0525 01:41:09.213289 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.12256 (* 0.0909091 = 0.283869 loss) | |
| I0525 01:41:09.213301 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.68687 (* 0.0909091 = 0.244261 loss) | |
| I0525 01:41:09.213315 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.15187 (* 0.0909091 = 0.286533 loss) | |
| I0525 01:41:09.213330 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 3.3449 (* 0.0909091 = 0.304082 loss) | |
| I0525 01:41:09.213343 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 2.55492 (* 0.0909091 = 0.232266 loss) | |
| I0525 01:41:09.213353 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00810991 (* 0.0909091 = 0.000737264 loss) | |
| I0525 01:41:09.213362 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00804949 (* 0.0909091 = 0.000731772 loss) | |
| I0525 01:41:09.213377 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00493125 (* 0.0909091 = 0.000448296 loss) | |
| I0525 01:41:09.213392 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00528493 (* 0.0909091 = 0.000480448 loss) | |
| I0525 01:41:09.213404 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00546518 (* 0.0909091 = 0.000496835 loss) | |
| I0525 01:41:09.213418 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00435257 (* 0.0909091 = 0.000395688 loss) | |
| I0525 01:41:09.213433 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0031808 (* 0.0909091 = 0.000289164 loss) | |
| I0525 01:41:09.213445 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00174619 (* 0.0909091 = 0.000158745 loss) | |
| I0525 01:41:09.213459 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00248791 (* 0.0909091 = 0.000226174 loss) | |
| I0525 01:41:09.213472 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00136565 (* 0.0909091 = 0.00012415 loss) | |
| I0525 01:41:09.213486 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000859398 (* 0.0909091 = 7.81271e-05 loss) | |
| I0525 01:41:09.213500 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000562007 (* 0.0909091 = 5.10916e-05 loss) | |
| I0525 01:41:09.213515 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000412822 (* 0.0909091 = 3.75292e-05 loss) | |
| I0525 01:41:09.213527 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000325635 (* 0.0909091 = 2.96032e-05 loss) | |
| I0525 01:41:09.213539 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:41:09.213551 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:41:09.213562 5272 solver.cpp:245] Train net output #149: total_confidence = 2.85074e-06 | |
| I0525 01:41:09.213584 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000257008 | |
| I0525 01:41:09.213598 5272 sgd_solver.cpp:106] Iteration 7000, lr = 0.001 | |
| I0525 01:42:31.897976 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 54.6448 > 30) by scale factor 0.549 | |
| I0525 01:44:15.763684 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.1021 > 30) by scale factor 0.74809 | |
| I0525 01:47:34.059463 5272 solver.cpp:229] Iteration 7500, loss = 11.139 | |
| I0525 01:47:34.059594 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.107143 | |
| I0525 01:47:34.059615 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 01:47:34.059628 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 01:47:34.059641 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 01:47:34.059654 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 01:47:34.059666 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0525 01:47:34.059679 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0525 01:47:34.059690 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 01:47:34.059702 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 01:47:34.059715 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 01:47:34.059727 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 01:47:34.059739 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 01:47:34.059751 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 01:47:34.059763 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 01:47:34.059774 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 01:47:34.059787 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 01:47:34.059798 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 01:47:34.059810 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 01:47:34.059823 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 01:47:34.059834 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:47:34.059846 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:47:34.059857 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:47:34.059870 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:47:34.059883 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.698864 | |
| I0525 01:47:34.059896 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.196429 | |
| I0525 01:47:34.059912 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.70661 (* 0.3 = 1.11198 loss) | |
| I0525 01:47:34.059926 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.31571 (* 0.3 = 0.394713 loss) | |
| I0525 01:47:34.059940 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.83821 (* 0.0272727 = 0.104679 loss) | |
| I0525 01:47:34.059954 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.47871 (* 0.0272727 = 0.094874 loss) | |
| I0525 01:47:34.059968 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.73681 (* 0.0272727 = 0.101913 loss) | |
| I0525 01:47:34.059981 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 4.05485 (* 0.0272727 = 0.110587 loss) | |
| I0525 01:47:34.059995 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 4.32254 (* 0.0272727 = 0.117888 loss) | |
| I0525 01:47:34.060009 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 4.51899 (* 0.0272727 = 0.123245 loss) | |
| I0525 01:47:34.060022 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.38146 (* 0.0272727 = 0.0649489 loss) | |
| I0525 01:47:34.060036 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.50198 (* 0.0272727 = 0.040963 loss) | |
| I0525 01:47:34.060050 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.749451 (* 0.0272727 = 0.0204396 loss) | |
| I0525 01:47:34.060063 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.619563 (* 0.0272727 = 0.0168972 loss) | |
| I0525 01:47:34.060077 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.576313 (* 0.0272727 = 0.0157176 loss) | |
| I0525 01:47:34.060092 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0897589 (* 0.0272727 = 0.00244797 loss) | |
| I0525 01:47:34.060106 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0759434 (* 0.0272727 = 0.00207118 loss) | |
| I0525 01:47:34.060139 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0595898 (* 0.0272727 = 0.00162518 loss) | |
| I0525 01:47:34.060155 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0434796 (* 0.0272727 = 0.00118581 loss) | |
| I0525 01:47:34.060169 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0372218 (* 0.0272727 = 0.00101514 loss) | |
| I0525 01:47:34.060183 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0540745 (* 0.0272727 = 0.00147476 loss) | |
| I0525 01:47:34.060197 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0360407 (* 0.0272727 = 0.000982928 loss) | |
| I0525 01:47:34.060211 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0362794 (* 0.0272727 = 0.000989438 loss) | |
| I0525 01:47:34.060225 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0241073 (* 0.0272727 = 0.000657472 loss) | |
| I0525 01:47:34.060240 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0283435 (* 0.0272727 = 0.000773003 loss) | |
| I0525 01:47:34.060253 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.039574 (* 0.0272727 = 0.00107929 loss) | |
| I0525 01:47:34.060266 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0892857 | |
| I0525 01:47:34.060278 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 01:47:34.060289 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 01:47:34.060302 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 01:47:34.060313 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 01:47:34.060324 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0525 01:47:34.060340 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0525 01:47:34.060364 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 01:47:34.060384 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 01:47:34.060397 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 01:47:34.060410 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 01:47:34.060421 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 01:47:34.060432 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 01:47:34.060444 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 01:47:34.060456 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 01:47:34.060467 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 01:47:34.060478 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 01:47:34.060489 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 01:47:34.060500 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 01:47:34.060513 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:47:34.060523 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:47:34.060534 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:47:34.060546 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:47:34.060557 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.710227 | |
| I0525 01:47:34.060570 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.267857 | |
| I0525 01:47:34.060583 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.73421 (* 0.3 = 1.12026 loss) | |
| I0525 01:47:34.060597 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.22424 (* 0.3 = 0.367273 loss) | |
| I0525 01:47:34.060616 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.18054 (* 0.0272727 = 0.114015 loss) | |
| I0525 01:47:34.060631 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.27276 (* 0.0272727 = 0.0892572 loss) | |
| I0525 01:47:34.060657 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.53983 (* 0.0272727 = 0.096541 loss) | |
| I0525 01:47:34.060672 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.96197 (* 0.0272727 = 0.108054 loss) | |
| I0525 01:47:34.060685 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.97925 (* 0.0272727 = 0.108525 loss) | |
| I0525 01:47:34.060699 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 4.50442 (* 0.0272727 = 0.122848 loss) | |
| I0525 01:47:34.060713 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.64991 (* 0.0272727 = 0.0722704 loss) | |
| I0525 01:47:34.060726 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.7004 (* 0.0272727 = 0.0463747 loss) | |
| I0525 01:47:34.060739 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.707858 (* 0.0272727 = 0.0193052 loss) | |
| I0525 01:47:34.060753 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.5563 (* 0.0272727 = 0.0151718 loss) | |
| I0525 01:47:34.060767 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.878436 (* 0.0272727 = 0.0239573 loss) | |
| I0525 01:47:34.060781 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.109182 (* 0.0272727 = 0.00297768 loss) | |
| I0525 01:47:34.060796 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0460682 (* 0.0272727 = 0.00125641 loss) | |
| I0525 01:47:34.060808 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0536117 (* 0.0272727 = 0.00146214 loss) | |
| I0525 01:47:34.060822 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0390222 (* 0.0272727 = 0.00106424 loss) | |
| I0525 01:47:34.060837 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0453714 (* 0.0272727 = 0.0012374 loss) | |
| I0525 01:47:34.060850 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0338976 (* 0.0272727 = 0.000924479 loss) | |
| I0525 01:47:34.060864 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0419012 (* 0.0272727 = 0.00114276 loss) | |
| I0525 01:47:34.060878 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0316819 (* 0.0272727 = 0.000864053 loss) | |
| I0525 01:47:34.060892 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0328677 (* 0.0272727 = 0.000896391 loss) | |
| I0525 01:47:34.060905 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0127353 (* 0.0272727 = 0.000347327 loss) | |
| I0525 01:47:34.060920 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0248026 (* 0.0272727 = 0.000676434 loss) | |
| I0525 01:47:34.060933 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0892857 | |
| I0525 01:47:34.060946 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 01:47:34.060958 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:47:34.060969 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:47:34.060981 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 01:47:34.060992 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0 | |
| I0525 01:47:34.061003 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0525 01:47:34.061015 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 01:47:34.061027 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 01:47:34.061038 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 01:47:34.061050 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 01:47:34.061061 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 01:47:34.061074 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 01:47:34.061085 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 01:47:34.061096 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 01:47:34.061107 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 01:47:34.061131 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 01:47:34.061158 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 01:47:34.061172 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 01:47:34.061183 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:47:34.061195 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:47:34.061206 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:47:34.061218 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:47:34.061229 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.710227 | |
| I0525 01:47:34.061241 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.285714 | |
| I0525 01:47:34.061255 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.4727 (* 1 = 3.4727 loss) | |
| I0525 01:47:34.061269 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.14198 (* 1 = 1.14198 loss) | |
| I0525 01:47:34.061282 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.16414 (* 0.0909091 = 0.287649 loss) | |
| I0525 01:47:34.061296 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.0435 (* 0.0909091 = 0.276682 loss) | |
| I0525 01:47:34.061310 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.17354 (* 0.0909091 = 0.288504 loss) | |
| I0525 01:47:34.061323 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.99009 (* 0.0909091 = 0.362736 loss) | |
| I0525 01:47:34.061336 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.73101 (* 0.0909091 = 0.339183 loss) | |
| I0525 01:47:34.061350 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.84399 (* 0.0909091 = 0.349454 loss) | |
| I0525 01:47:34.061364 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.11593 (* 0.0909091 = 0.192357 loss) | |
| I0525 01:47:34.061374 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.85874 (* 0.0909091 = 0.168976 loss) | |
| I0525 01:47:34.061383 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.677261 (* 0.0909091 = 0.0615691 loss) | |
| I0525 01:47:34.061393 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.765279 (* 0.0909091 = 0.0695708 loss) | |
| I0525 01:47:34.061408 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.832122 (* 0.0909091 = 0.0756475 loss) | |
| I0525 01:47:34.061422 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0261005 (* 0.0909091 = 0.00237277 loss) | |
| I0525 01:47:34.061436 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0292434 (* 0.0909091 = 0.00265849 loss) | |
| I0525 01:47:34.061450 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0178083 (* 0.0909091 = 0.00161894 loss) | |
| I0525 01:47:34.061463 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0195543 (* 0.0909091 = 0.00177767 loss) | |
| I0525 01:47:34.061477 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0164549 (* 0.0909091 = 0.0014959 loss) | |
| I0525 01:47:34.061491 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0202718 (* 0.0909091 = 0.00184289 loss) | |
| I0525 01:47:34.061506 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0129819 (* 0.0909091 = 0.00118017 loss) | |
| I0525 01:47:34.061519 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0119746 (* 0.0909091 = 0.0010886 loss) | |
| I0525 01:47:34.061533 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00923158 (* 0.0909091 = 0.000839234 loss) | |
| I0525 01:47:34.061547 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00720708 (* 0.0909091 = 0.000655189 loss) | |
| I0525 01:47:34.061560 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0119367 (* 0.0909091 = 0.00108516 loss) | |
| I0525 01:47:34.061573 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:47:34.061584 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:47:34.061595 5272 solver.cpp:245] Train net output #149: total_confidence = 3.45977e-05 | |
| I0525 01:47:34.061616 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00125718 | |
| I0525 01:47:34.061631 5272 sgd_solver.cpp:106] Iteration 7500, lr = 0.001 | |
| I0525 01:49:56.031536 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.4072 > 30) by scale factor 0.761283 | |
| I0525 01:49:58.343458 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 56.848 > 30) by scale factor 0.527723 | |
| I0525 01:50:45.273576 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.1743 > 30) by scale factor 0.932423 | |
| I0525 01:52:23.794315 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.4524 > 30) by scale factor 0.741612 | |
| I0525 01:52:26.876194 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.9805 > 30) by scale factor 0.85762 | |
| I0525 01:53:58.859163 5272 solver.cpp:229] Iteration 8000, loss = 10.8864 | |
| I0525 01:53:58.859300 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0535714 | |
| I0525 01:53:58.859321 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 01:53:58.859335 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 01:53:58.859349 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 01:53:58.859361 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0525 01:53:58.859374 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 01:53:58.859385 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 01:53:58.859397 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 01:53:58.859410 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 01:53:58.859421 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0525 01:53:58.859433 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 01:53:58.859447 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 01:53:58.859458 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 01:53:58.859470 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 01:53:58.859483 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 01:53:58.859494 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 01:53:58.859506 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 01:53:58.859518 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0525 01:53:58.859529 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0525 01:53:58.859541 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 01:53:58.859554 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 01:53:58.859565 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 01:53:58.859576 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 01:53:58.859588 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0525 01:53:58.859601 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.214286 | |
| I0525 01:53:58.859616 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.78388 (* 0.3 = 1.13516 loss) | |
| I0525 01:53:58.859630 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.38411 (* 0.3 = 0.415233 loss) | |
| I0525 01:53:58.859645 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.373 (* 0.0272727 = 0.0919908 loss) | |
| I0525 01:53:58.859659 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.49751 (* 0.0272727 = 0.0953867 loss) | |
| I0525 01:53:58.859673 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.61831 (* 0.0272727 = 0.0986813 loss) | |
| I0525 01:53:58.859686 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.03267 (* 0.0272727 = 0.0827092 loss) | |
| I0525 01:53:58.859700 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.57303 (* 0.0272727 = 0.0701736 loss) | |
| I0525 01:53:58.859714 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.18166 (* 0.0272727 = 0.0594998 loss) | |
| I0525 01:53:58.859729 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.60866 (* 0.0272727 = 0.0711453 loss) | |
| I0525 01:53:58.859741 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.40907 (* 0.0272727 = 0.0384292 loss) | |
| I0525 01:53:58.859755 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.16809 (* 0.0272727 = 0.0318569 loss) | |
| I0525 01:53:58.859769 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.876062 (* 0.0272727 = 0.0238926 loss) | |
| I0525 01:53:58.859783 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.771861 (* 0.0272727 = 0.0210508 loss) | |
| I0525 01:53:58.859797 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 1.11626 (* 0.0272727 = 0.0304434 loss) | |
| I0525 01:53:58.859832 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 1.14765 (* 0.0272727 = 0.0312997 loss) | |
| I0525 01:53:58.859846 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 1.04524 (* 0.0272727 = 0.0285065 loss) | |
| I0525 01:53:58.859860 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 1.20448 (* 0.0272727 = 0.0328496 loss) | |
| I0525 01:53:58.859877 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.88867 (* 0.0272727 = 0.0242364 loss) | |
| I0525 01:53:58.859892 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 1.04917 (* 0.0272727 = 0.0286137 loss) | |
| I0525 01:53:58.859906 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 1.03483 (* 0.0272727 = 0.0282227 loss) | |
| I0525 01:53:58.859920 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00314976 (* 0.0272727 = 8.59024e-05 loss) | |
| I0525 01:53:58.859935 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00195826 (* 0.0272727 = 5.34072e-05 loss) | |
| I0525 01:53:58.859948 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0019175 (* 0.0272727 = 5.22955e-05 loss) | |
| I0525 01:53:58.859962 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00178133 (* 0.0272727 = 4.85817e-05 loss) | |
| I0525 01:53:58.859975 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0892857 | |
| I0525 01:53:58.859987 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 01:53:58.859999 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 01:53:58.860011 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 01:53:58.860023 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 01:53:58.860034 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0525 01:53:58.860045 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 01:53:58.860057 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 01:53:58.860069 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 01:53:58.860080 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0525 01:53:58.860092 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 01:53:58.860105 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 01:53:58.860116 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 01:53:58.860127 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 01:53:58.860138 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 01:53:58.860151 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 01:53:58.860162 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 01:53:58.860173 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0525 01:53:58.860185 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0525 01:53:58.860198 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 01:53:58.860208 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 01:53:58.860220 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 01:53:58.860229 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 01:53:58.860235 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0525 01:53:58.860249 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0525 01:53:58.860262 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.64118 (* 0.3 = 1.09236 loss) | |
| I0525 01:53:58.860276 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.36725 (* 0.3 = 0.410175 loss) | |
| I0525 01:53:58.860291 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.29495 (* 0.0272727 = 0.0898624 loss) | |
| I0525 01:53:58.860304 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.51058 (* 0.0272727 = 0.123016 loss) | |
| I0525 01:53:58.860333 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.92163 (* 0.0272727 = 0.106954 loss) | |
| I0525 01:53:58.860348 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.13927 (* 0.0272727 = 0.0856165 loss) | |
| I0525 01:53:58.860363 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.73348 (* 0.0272727 = 0.0745494 loss) | |
| I0525 01:53:58.860375 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.41893 (* 0.0272727 = 0.0659707 loss) | |
| I0525 01:53:58.860389 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.39883 (* 0.0272727 = 0.0654227 loss) | |
| I0525 01:53:58.860402 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.36857 (* 0.0272727 = 0.0373246 loss) | |
| I0525 01:53:58.860416 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.950654 (* 0.0272727 = 0.0259269 loss) | |
| I0525 01:53:58.860430 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.921312 (* 0.0272727 = 0.0251267 loss) | |
| I0525 01:53:58.860443 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.952763 (* 0.0272727 = 0.0259844 loss) | |
| I0525 01:53:58.860457 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 1.04418 (* 0.0272727 = 0.0284776 loss) | |
| I0525 01:53:58.860471 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 1.11922 (* 0.0272727 = 0.0305242 loss) | |
| I0525 01:53:58.860484 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.85363 (* 0.0272727 = 0.0232808 loss) | |
| I0525 01:53:58.860498 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 1.09577 (* 0.0272727 = 0.0298846 loss) | |
| I0525 01:53:58.860512 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 1.18136 (* 0.0272727 = 0.0322188 loss) | |
| I0525 01:53:58.860525 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.930875 (* 0.0272727 = 0.0253875 loss) | |
| I0525 01:53:58.860539 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 1.17568 (* 0.0272727 = 0.032064 loss) | |
| I0525 01:53:58.860553 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0240429 (* 0.0272727 = 0.000655717 loss) | |
| I0525 01:53:58.860566 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00897186 (* 0.0272727 = 0.000244687 loss) | |
| I0525 01:53:58.860580 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00976135 (* 0.0272727 = 0.000266219 loss) | |
| I0525 01:53:58.860594 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.015623 (* 0.0272727 = 0.000426082 loss) | |
| I0525 01:53:58.860606 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.125 | |
| I0525 01:53:58.860618 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 01:53:58.860630 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 01:53:58.860641 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 01:53:58.860653 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 01:53:58.860664 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 01:53:58.860677 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 01:53:58.860688 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 01:53:58.860699 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 01:53:58.860712 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0525 01:53:58.860723 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 01:53:58.860733 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 01:53:58.860745 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 01:53:58.860757 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 01:53:58.860769 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 01:53:58.860780 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 01:53:58.860792 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 01:53:58.860813 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0525 01:53:58.860826 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0525 01:53:58.860838 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 01:53:58.860851 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 01:53:58.860862 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 01:53:58.860873 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 01:53:58.860885 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0525 01:53:58.860896 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.339286 | |
| I0525 01:53:58.860910 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.37032 (* 1 = 3.37032 loss) | |
| I0525 01:53:58.860924 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.33828 (* 1 = 1.33828 loss) | |
| I0525 01:53:58.860941 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.16327 (* 0.0909091 = 0.28757 loss) | |
| I0525 01:53:58.860955 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.29743 (* 0.0909091 = 0.299767 loss) | |
| I0525 01:53:58.860970 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.34765 (* 0.0909091 = 0.304332 loss) | |
| I0525 01:53:58.860982 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.92708 (* 0.0909091 = 0.266098 loss) | |
| I0525 01:53:58.860996 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.5584 (* 0.0909091 = 0.232582 loss) | |
| I0525 01:53:58.861006 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.10811 (* 0.0909091 = 0.191647 loss) | |
| I0525 01:53:58.861016 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.13003 (* 0.0909091 = 0.193639 loss) | |
| I0525 01:53:58.861029 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.21617 (* 0.0909091 = 0.110561 loss) | |
| I0525 01:53:58.861043 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.898435 (* 0.0909091 = 0.0816759 loss) | |
| I0525 01:53:58.861057 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.850684 (* 0.0909091 = 0.0773349 loss) | |
| I0525 01:53:58.861071 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.82365 (* 0.0909091 = 0.0748773 loss) | |
| I0525 01:53:58.861085 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 1.0669 (* 0.0909091 = 0.0969913 loss) | |
| I0525 01:53:58.861099 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.821967 (* 0.0909091 = 0.0747242 loss) | |
| I0525 01:53:58.861112 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.759522 (* 0.0909091 = 0.0690475 loss) | |
| I0525 01:53:58.861140 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 1.13005 (* 0.0909091 = 0.102732 loss) | |
| I0525 01:53:58.861156 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.68099 (* 0.0909091 = 0.0619082 loss) | |
| I0525 01:53:58.861169 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.913378 (* 0.0909091 = 0.0830344 loss) | |
| I0525 01:53:58.861183 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 1.07236 (* 0.0909091 = 0.0974871 loss) | |
| I0525 01:53:58.861197 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00341045 (* 0.0909091 = 0.000310041 loss) | |
| I0525 01:53:58.861212 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00107248 (* 0.0909091 = 9.74978e-05 loss) | |
| I0525 01:53:58.861225 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00075675 (* 0.0909091 = 6.87954e-05 loss) | |
| I0525 01:53:58.861238 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000565158 (* 0.0909091 = 5.1378e-05 loss) | |
| I0525 01:53:58.861250 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 01:53:58.861263 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 01:53:58.861274 5272 solver.cpp:245] Train net output #149: total_confidence = 4.91878e-06 | |
| I0525 01:53:58.861295 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 9.80873e-05 | |
| I0525 01:53:58.861310 5272 sgd_solver.cpp:106] Iteration 8000, lr = 0.001 | |
| I0525 01:54:09.996558 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.7448 > 30) by scale factor 0.916177 | |
| I0525 01:54:46.928495 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.3652 > 30) by scale factor 0.743215 | |
| I0525 01:55:52.398910 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.2681 > 30) by scale factor 0.827173 | |
| I0525 01:57:37.858790 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.8314 > 30) by scale factor 0.913758 | |
| I0525 01:59:34.024374 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 46.7965 > 30) by scale factor 0.641073 | |
| I0525 02:00:23.710636 5272 solver.cpp:229] Iteration 8500, loss = 10.7341 | |
| I0525 02:00:23.710764 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0525 02:00:23.710784 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 02:00:23.710796 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 02:00:23.710809 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 02:00:23.710821 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 02:00:23.710834 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 02:00:23.710846 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 02:00:23.710858 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 02:00:23.710870 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 02:00:23.710886 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 02:00:23.710898 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 02:00:23.710911 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 02:00:23.710922 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 02:00:23.710933 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 02:00:23.710945 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:00:23.710958 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:00:23.710968 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:00:23.710980 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:00:23.710993 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:00:23.711004 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:00:23.711015 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:00:23.711027 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:00:23.711040 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:00:23.711050 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0525 02:00:23.711062 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.136364 | |
| I0525 02:00:23.711078 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.45275 (* 0.3 = 1.03583 loss) | |
| I0525 02:00:23.711092 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.05491 (* 0.3 = 0.316473 loss) | |
| I0525 02:00:23.711107 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.93602 (* 0.0272727 = 0.107346 loss) | |
| I0525 02:00:23.711120 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.72846 (* 0.0272727 = 0.101685 loss) | |
| I0525 02:00:23.711134 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.60583 (* 0.0272727 = 0.0983408 loss) | |
| I0525 02:00:23.711148 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.60037 (* 0.0272727 = 0.0981919 loss) | |
| I0525 02:00:23.711163 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.79114 (* 0.0272727 = 0.076122 loss) | |
| I0525 02:00:23.711176 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.31816 (* 0.0272727 = 0.0632224 loss) | |
| I0525 02:00:23.711189 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.31524 (* 0.0272727 = 0.0358701 loss) | |
| I0525 02:00:23.711204 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.968286 (* 0.0272727 = 0.0264078 loss) | |
| I0525 02:00:23.711218 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.188955 (* 0.0272727 = 0.00515331 loss) | |
| I0525 02:00:23.711232 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.197029 (* 0.0272727 = 0.00537352 loss) | |
| I0525 02:00:23.711246 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.11399 (* 0.0272727 = 0.00310882 loss) | |
| I0525 02:00:23.711261 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.141564 (* 0.0272727 = 0.00386083 loss) | |
| I0525 02:00:23.711274 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0633773 (* 0.0272727 = 0.00172847 loss) | |
| I0525 02:00:23.711308 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0542592 (* 0.0272727 = 0.0014798 loss) | |
| I0525 02:00:23.711324 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0639472 (* 0.0272727 = 0.00174401 loss) | |
| I0525 02:00:23.711338 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0540718 (* 0.0272727 = 0.00147468 loss) | |
| I0525 02:00:23.711352 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0362595 (* 0.0272727 = 0.000988897 loss) | |
| I0525 02:00:23.711366 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0346971 (* 0.0272727 = 0.000946286 loss) | |
| I0525 02:00:23.711380 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0258803 (* 0.0272727 = 0.000705826 loss) | |
| I0525 02:00:23.711395 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.04423 (* 0.0272727 = 0.00120627 loss) | |
| I0525 02:00:23.711408 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0121073 (* 0.0272727 = 0.0003302 loss) | |
| I0525 02:00:23.711422 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0140672 (* 0.0272727 = 0.000383652 loss) | |
| I0525 02:00:23.711434 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0681818 | |
| I0525 02:00:23.711447 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.375 | |
| I0525 02:00:23.711458 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 02:00:23.711469 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 02:00:23.711480 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 02:00:23.711493 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 02:00:23.711504 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 02:00:23.711516 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 02:00:23.711527 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 02:00:23.711540 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 02:00:23.711551 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 02:00:23.711562 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 02:00:23.711573 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 02:00:23.711585 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 02:00:23.711596 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:00:23.711607 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:00:23.711619 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:00:23.711630 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:00:23.711642 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:00:23.711653 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:00:23.711664 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:00:23.711676 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:00:23.711688 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:00:23.711699 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0525 02:00:23.711710 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.136364 | |
| I0525 02:00:23.711724 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.31637 (* 0.3 = 0.994912 loss) | |
| I0525 02:00:23.711737 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.07234 (* 0.3 = 0.321701 loss) | |
| I0525 02:00:23.711751 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.76523 (* 0.0272727 = 0.0754154 loss) | |
| I0525 02:00:23.711765 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.02498 (* 0.0272727 = 0.109772 loss) | |
| I0525 02:00:23.711778 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.68976 (* 0.0272727 = 0.10063 loss) | |
| I0525 02:00:23.711807 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.9057 (* 0.0272727 = 0.106519 loss) | |
| I0525 02:00:23.711822 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.88556 (* 0.0272727 = 0.0786971 loss) | |
| I0525 02:00:23.711836 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.20065 (* 0.0272727 = 0.0600176 loss) | |
| I0525 02:00:23.711849 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.51796 (* 0.0272727 = 0.0413988 loss) | |
| I0525 02:00:23.711863 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.14035 (* 0.0272727 = 0.0311004 loss) | |
| I0525 02:00:23.711877 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0608464 (* 0.0272727 = 0.00165945 loss) | |
| I0525 02:00:23.711891 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0461119 (* 0.0272727 = 0.0012576 loss) | |
| I0525 02:00:23.711905 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0513955 (* 0.0272727 = 0.0014017 loss) | |
| I0525 02:00:23.711920 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0191378 (* 0.0272727 = 0.000521939 loss) | |
| I0525 02:00:23.711936 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0133266 (* 0.0272727 = 0.000363453 loss) | |
| I0525 02:00:23.711951 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0111529 (* 0.0272727 = 0.000304169 loss) | |
| I0525 02:00:23.711964 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0133932 (* 0.0272727 = 0.000365268 loss) | |
| I0525 02:00:23.711979 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00262937 (* 0.0272727 = 7.171e-05 loss) | |
| I0525 02:00:23.711993 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00284424 (* 0.0272727 = 7.75701e-05 loss) | |
| I0525 02:00:23.712007 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00184793 (* 0.0272727 = 5.0398e-05 loss) | |
| I0525 02:00:23.712020 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00189623 (* 0.0272727 = 5.17154e-05 loss) | |
| I0525 02:00:23.712034 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00187828 (* 0.0272727 = 5.12257e-05 loss) | |
| I0525 02:00:23.712049 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00186272 (* 0.0272727 = 5.08015e-05 loss) | |
| I0525 02:00:23.712062 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00276296 (* 0.0272727 = 7.53534e-05 loss) | |
| I0525 02:00:23.712074 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0227273 | |
| I0525 02:00:23.712085 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 02:00:23.712097 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 02:00:23.712105 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 02:00:23.712113 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 02:00:23.712126 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 02:00:23.712138 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 02:00:23.712149 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 02:00:23.712162 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 02:00:23.712173 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 02:00:23.712185 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 02:00:23.712196 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 02:00:23.712208 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 02:00:23.712219 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 02:00:23.712230 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:00:23.712242 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:00:23.712254 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:00:23.712275 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:00:23.712288 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:00:23.712299 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:00:23.712311 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:00:23.712322 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:00:23.712334 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:00:23.712345 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0525 02:00:23.712357 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.295455 | |
| I0525 02:00:23.712370 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.20822 (* 1 = 3.20822 loss) | |
| I0525 02:00:23.712384 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.925512 (* 1 = 0.925512 loss) | |
| I0525 02:00:23.712399 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.16606 (* 0.0909091 = 0.287824 loss) | |
| I0525 02:00:23.712412 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.30318 (* 0.0909091 = 0.300289 loss) | |
| I0525 02:00:23.712425 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.96121 (* 0.0909091 = 0.269201 loss) | |
| I0525 02:00:23.712440 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.53237 (* 0.0909091 = 0.321125 loss) | |
| I0525 02:00:23.712452 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.83622 (* 0.0909091 = 0.257838 loss) | |
| I0525 02:00:23.712466 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.038 (* 0.0909091 = 0.185273 loss) | |
| I0525 02:00:23.712479 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.01494 (* 0.0909091 = 0.0922673 loss) | |
| I0525 02:00:23.712493 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.8627 (* 0.0909091 = 0.0784273 loss) | |
| I0525 02:00:23.712507 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0755781 (* 0.0909091 = 0.00687073 loss) | |
| I0525 02:00:23.712522 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0220414 (* 0.0909091 = 0.00200377 loss) | |
| I0525 02:00:23.712534 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0163476 (* 0.0909091 = 0.00148615 loss) | |
| I0525 02:00:23.712548 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0169783 (* 0.0909091 = 0.00154349 loss) | |
| I0525 02:00:23.712563 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00921395 (* 0.0909091 = 0.000837632 loss) | |
| I0525 02:00:23.712576 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0076564 (* 0.0909091 = 0.000696037 loss) | |
| I0525 02:00:23.712589 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00792762 (* 0.0909091 = 0.000720693 loss) | |
| I0525 02:00:23.712604 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00583303 (* 0.0909091 = 0.000530276 loss) | |
| I0525 02:00:23.712616 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00602691 (* 0.0909091 = 0.000547901 loss) | |
| I0525 02:00:23.712630 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00258678 (* 0.0909091 = 0.000235161 loss) | |
| I0525 02:00:23.712644 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00198702 (* 0.0909091 = 0.000180638 loss) | |
| I0525 02:00:23.712658 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00184639 (* 0.0909091 = 0.000167853 loss) | |
| I0525 02:00:23.712672 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000865449 (* 0.0909091 = 7.86771e-05 loss) | |
| I0525 02:00:23.712685 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000791186 (* 0.0909091 = 7.1926e-05 loss) | |
| I0525 02:00:23.712697 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:00:23.712709 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:00:23.712720 5272 solver.cpp:245] Train net output #149: total_confidence = 0.00010174 | |
| I0525 02:00:23.712741 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000433083 | |
| I0525 02:00:23.712755 5272 sgd_solver.cpp:106] Iteration 8500, lr = 0.001 | |
| I0525 02:03:41.112161 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8669 > 30) by scale factor 0.971915 | |
| I0525 02:06:48.461539 5272 solver.cpp:229] Iteration 9000, loss = 10.6014 | |
| I0525 02:06:48.461696 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0454545 | |
| I0525 02:06:48.461717 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 02:06:48.461731 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 02:06:48.461745 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 02:06:48.461757 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 02:06:48.461769 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 02:06:48.461781 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 02:06:48.461793 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 02:06:48.461805 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 02:06:48.461818 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 02:06:48.461829 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 02:06:48.461841 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 02:06:48.461853 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 02:06:48.461865 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 02:06:48.461879 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:06:48.461892 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:06:48.461905 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:06:48.461916 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:06:48.461928 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:06:48.461941 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:06:48.461952 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:06:48.461964 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:06:48.461976 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:06:48.461987 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.761364 | |
| I0525 02:06:48.461999 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.204545 | |
| I0525 02:06:48.462015 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.2905 (* 0.3 = 0.987151 loss) | |
| I0525 02:06:48.462030 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.934881 (* 0.3 = 0.280464 loss) | |
| I0525 02:06:48.462044 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.92566 (* 0.0272727 = 0.0797907 loss) | |
| I0525 02:06:48.462064 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.41233 (* 0.0272727 = 0.0930636 loss) | |
| I0525 02:06:48.462091 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.63043 (* 0.0272727 = 0.0990117 loss) | |
| I0525 02:06:48.462118 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.81988 (* 0.0272727 = 0.104178 loss) | |
| I0525 02:06:48.462144 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.91326 (* 0.0272727 = 0.0794526 loss) | |
| I0525 02:06:48.462167 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.6599 (* 0.0272727 = 0.0725428 loss) | |
| I0525 02:06:48.462189 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.10958 (* 0.0272727 = 0.0302613 loss) | |
| I0525 02:06:48.462213 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.02383 (* 0.0272727 = 0.0279226 loss) | |
| I0525 02:06:48.462235 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0274324 (* 0.0272727 = 0.000748155 loss) | |
| I0525 02:06:48.462260 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0283349 (* 0.0272727 = 0.000772769 loss) | |
| I0525 02:06:48.462281 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0243095 (* 0.0272727 = 0.000662988 loss) | |
| I0525 02:06:48.462303 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0184351 (* 0.0272727 = 0.000502777 loss) | |
| I0525 02:06:48.462327 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0127289 (* 0.0272727 = 0.000347152 loss) | |
| I0525 02:06:48.462383 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0173868 (* 0.0272727 = 0.000474186 loss) | |
| I0525 02:06:48.462409 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0104953 (* 0.0272727 = 0.000286234 loss) | |
| I0525 02:06:48.462435 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0161775 (* 0.0272727 = 0.000441205 loss) | |
| I0525 02:06:48.462461 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.013427 (* 0.0272727 = 0.000366191 loss) | |
| I0525 02:06:48.462486 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0124264 (* 0.0272727 = 0.000338902 loss) | |
| I0525 02:06:48.462512 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.012312 (* 0.0272727 = 0.00033578 loss) | |
| I0525 02:06:48.462538 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0143283 (* 0.0272727 = 0.000390773 loss) | |
| I0525 02:06:48.462565 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0160374 (* 0.0272727 = 0.000437385 loss) | |
| I0525 02:06:48.462591 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0091108 (* 0.0272727 = 0.000248476 loss) | |
| I0525 02:06:48.462613 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0909091 | |
| I0525 02:06:48.462635 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 02:06:48.462656 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 02:06:48.462676 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 02:06:48.462697 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 02:06:48.462723 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 02:06:48.462745 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 02:06:48.462767 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 02:06:48.462788 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 02:06:48.462810 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 02:06:48.462831 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 02:06:48.462852 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 02:06:48.462873 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 02:06:48.462896 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 02:06:48.462918 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:06:48.462944 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:06:48.462965 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:06:48.462986 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:06:48.463007 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:06:48.463029 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:06:48.463052 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:06:48.463073 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:06:48.463095 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:06:48.463116 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0525 02:06:48.463138 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.227273 | |
| I0525 02:06:48.463165 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.21606 (* 0.3 = 0.964817 loss) | |
| I0525 02:06:48.463191 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.0471 (* 0.3 = 0.314131 loss) | |
| I0525 02:06:48.463217 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.31807 (* 0.0272727 = 0.0904929 loss) | |
| I0525 02:06:48.463240 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.58032 (* 0.0272727 = 0.0976452 loss) | |
| I0525 02:06:48.463282 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.74134 (* 0.0272727 = 0.102037 loss) | |
| I0525 02:06:48.463310 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.37695 (* 0.0272727 = 0.0920986 loss) | |
| I0525 02:06:48.463335 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.30345 (* 0.0272727 = 0.090094 loss) | |
| I0525 02:06:48.463358 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.51754 (* 0.0272727 = 0.0686602 loss) | |
| I0525 02:06:48.463382 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.24359 (* 0.0272727 = 0.0339161 loss) | |
| I0525 02:06:48.463407 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.14515 (* 0.0272727 = 0.0312312 loss) | |
| I0525 02:06:48.463435 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0495629 (* 0.0272727 = 0.00135171 loss) | |
| I0525 02:06:48.463460 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0401712 (* 0.0272727 = 0.00109558 loss) | |
| I0525 02:06:48.463485 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0386994 (* 0.0272727 = 0.00105544 loss) | |
| I0525 02:06:48.463510 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0203508 (* 0.0272727 = 0.000555022 loss) | |
| I0525 02:06:48.463536 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0341986 (* 0.0272727 = 0.000932689 loss) | |
| I0525 02:06:48.463562 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0336644 (* 0.0272727 = 0.00091812 loss) | |
| I0525 02:06:48.463587 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0176802 (* 0.0272727 = 0.000482186 loss) | |
| I0525 02:06:48.463610 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0108395 (* 0.0272727 = 0.000295623 loss) | |
| I0525 02:06:48.463636 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00564106 (* 0.0272727 = 0.000153847 loss) | |
| I0525 02:06:48.463662 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00481375 (* 0.0272727 = 0.000131284 loss) | |
| I0525 02:06:48.463690 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00289966 (* 0.0272727 = 7.90816e-05 loss) | |
| I0525 02:06:48.463716 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00364745 (* 0.0272727 = 9.94759e-05 loss) | |
| I0525 02:06:48.463739 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00319706 (* 0.0272727 = 8.71926e-05 loss) | |
| I0525 02:06:48.463769 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00267019 (* 0.0272727 = 7.28233e-05 loss) | |
| I0525 02:06:48.463791 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0681818 | |
| I0525 02:06:48.463810 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0525 02:06:48.463831 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0525 02:06:48.463855 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 02:06:48.463876 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 02:06:48.463896 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 02:06:48.463917 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 02:06:48.463939 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 02:06:48.463958 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 02:06:48.463971 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 02:06:48.463985 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 02:06:48.463997 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 02:06:48.464010 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 02:06:48.464020 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 02:06:48.464032 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:06:48.464043 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:06:48.464056 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:06:48.464079 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:06:48.464092 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:06:48.464104 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:06:48.464117 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:06:48.464128 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:06:48.464139 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:06:48.464151 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.761364 | |
| I0525 02:06:48.464164 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.272727 | |
| I0525 02:06:48.464179 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.14527 (* 1 = 3.14527 loss) | |
| I0525 02:06:48.464191 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.865455 (* 1 = 0.865455 loss) | |
| I0525 02:06:48.464205 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.75123 (* 0.0909091 = 0.250112 loss) | |
| I0525 02:06:48.464220 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.88901 (* 0.0909091 = 0.262637 loss) | |
| I0525 02:06:48.464233 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.34774 (* 0.0909091 = 0.30434 loss) | |
| I0525 02:06:48.464247 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.52955 (* 0.0909091 = 0.320868 loss) | |
| I0525 02:06:48.464260 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.76007 (* 0.0909091 = 0.250915 loss) | |
| I0525 02:06:48.464274 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.13938 (* 0.0909091 = 0.194489 loss) | |
| I0525 02:06:48.464288 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.639256 (* 0.0909091 = 0.0581142 loss) | |
| I0525 02:06:48.464303 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.577519 (* 0.0909091 = 0.0525018 loss) | |
| I0525 02:06:48.464316 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0470034 (* 0.0909091 = 0.00427304 loss) | |
| I0525 02:06:48.464330 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.02836 (* 0.0909091 = 0.00257818 loss) | |
| I0525 02:06:48.464344 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0219587 (* 0.0909091 = 0.00199625 loss) | |
| I0525 02:06:48.464359 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00828307 (* 0.0909091 = 0.000753007 loss) | |
| I0525 02:06:48.464371 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0177632 (* 0.0909091 = 0.00161483 loss) | |
| I0525 02:06:48.464386 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0114985 (* 0.0909091 = 0.00104532 loss) | |
| I0525 02:06:48.464401 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00804281 (* 0.0909091 = 0.000731165 loss) | |
| I0525 02:06:48.464413 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0052305 (* 0.0909091 = 0.0004755 loss) | |
| I0525 02:06:48.464428 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00891069 (* 0.0909091 = 0.000810063 loss) | |
| I0525 02:06:48.464442 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00359177 (* 0.0909091 = 0.000326525 loss) | |
| I0525 02:06:48.464455 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00278682 (* 0.0909091 = 0.000253347 loss) | |
| I0525 02:06:48.464469 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00379981 (* 0.0909091 = 0.000345437 loss) | |
| I0525 02:06:48.464483 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00323816 (* 0.0909091 = 0.000294379 loss) | |
| I0525 02:06:48.464498 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00219976 (* 0.0909091 = 0.000199978 loss) | |
| I0525 02:06:48.464509 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:06:48.464521 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:06:48.464532 5272 solver.cpp:245] Train net output #149: total_confidence = 1.59605e-05 | |
| I0525 02:06:48.464553 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000424533 | |
| I0525 02:06:48.464568 5272 sgd_solver.cpp:106] Iteration 9000, lr = 0.001 | |
| I0525 02:07:58.845538 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 39.6896 > 30) by scale factor 0.755866 | |
| I0525 02:10:41.983757 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.8368 > 30) by scale factor 0.684356 | |
| I0525 02:13:13.249094 5272 solver.cpp:229] Iteration 9500, loss = 10.592 | |
| I0525 02:13:13.249265 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0425532 | |
| I0525 02:13:13.249286 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 02:13:13.249300 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 02:13:13.249312 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 02:13:13.249325 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 02:13:13.249336 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 02:13:13.249348 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0525 02:13:13.249361 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 02:13:13.249373 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 02:13:13.249385 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 02:13:13.249397 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 02:13:13.249409 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 02:13:13.249421 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 02:13:13.249433 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 02:13:13.249445 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:13:13.249456 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:13:13.249469 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:13:13.249480 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:13:13.249492 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:13:13.249505 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:13:13.249516 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:13:13.249528 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:13:13.249539 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:13:13.249552 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0525 02:13:13.249563 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.0851064 | |
| I0525 02:13:13.249580 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.58935 (* 0.3 = 1.07681 loss) | |
| I0525 02:13:13.249595 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.0816 (* 0.3 = 0.324479 loss) | |
| I0525 02:13:13.249609 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.16306 (* 0.0272727 = 0.0862652 loss) | |
| I0525 02:13:13.249624 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.11088 (* 0.0272727 = 0.112115 loss) | |
| I0525 02:13:13.249637 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.5617 (* 0.0272727 = 0.0971373 loss) | |
| I0525 02:13:13.249651 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.35086 (* 0.0272727 = 0.091387 loss) | |
| I0525 02:13:13.249666 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.25701 (* 0.0272727 = 0.0888276 loss) | |
| I0525 02:13:13.249681 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 4.00934 (* 0.0272727 = 0.109346 loss) | |
| I0525 02:13:13.249696 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.56966 (* 0.0272727 = 0.0428089 loss) | |
| I0525 02:13:13.249709 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.65934 (* 0.0272727 = 0.017982 loss) | |
| I0525 02:13:13.249723 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.340386 (* 0.0272727 = 0.00928325 loss) | |
| I0525 02:13:13.249737 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.257514 (* 0.0272727 = 0.00702312 loss) | |
| I0525 02:13:13.249752 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.226935 (* 0.0272727 = 0.00618915 loss) | |
| I0525 02:13:13.249766 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.142721 (* 0.0272727 = 0.00389238 loss) | |
| I0525 02:13:13.249779 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.134003 (* 0.0272727 = 0.00365462 loss) | |
| I0525 02:13:13.249816 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0651643 (* 0.0272727 = 0.00177721 loss) | |
| I0525 02:13:13.249831 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0500841 (* 0.0272727 = 0.00136593 loss) | |
| I0525 02:13:13.249846 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0616478 (* 0.0272727 = 0.0016813 loss) | |
| I0525 02:13:13.249861 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0238216 (* 0.0272727 = 0.000649681 loss) | |
| I0525 02:13:13.249878 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0367692 (* 0.0272727 = 0.0010028 loss) | |
| I0525 02:13:13.249893 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0191323 (* 0.0272727 = 0.000521791 loss) | |
| I0525 02:13:13.249908 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00723268 (* 0.0272727 = 0.000197255 loss) | |
| I0525 02:13:13.249922 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0132136 (* 0.0272727 = 0.000360371 loss) | |
| I0525 02:13:13.249936 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0114148 (* 0.0272727 = 0.000311312 loss) | |
| I0525 02:13:13.249948 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0425532 | |
| I0525 02:13:13.249960 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 02:13:13.249972 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 02:13:13.249984 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0525 02:13:13.249995 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 02:13:13.250007 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 02:13:13.250020 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 02:13:13.250031 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 02:13:13.250043 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 02:13:13.250056 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 02:13:13.250066 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 02:13:13.250078 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 02:13:13.250089 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 02:13:13.250102 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 02:13:13.250113 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:13:13.250124 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:13:13.250135 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:13:13.250149 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:13:13.250160 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:13:13.250171 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:13:13.250183 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:13:13.250195 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:13:13.250206 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:13:13.250217 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.698864 | |
| I0525 02:13:13.250229 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.106383 | |
| I0525 02:13:13.250243 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.71447 (* 0.3 = 1.11434 loss) | |
| I0525 02:13:13.250257 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.2879 (* 0.3 = 0.386371 loss) | |
| I0525 02:13:13.250272 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.6602 (* 0.0272727 = 0.0998235 loss) | |
| I0525 02:13:13.250290 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.88976 (* 0.0272727 = 0.106084 loss) | |
| I0525 02:13:13.250304 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.45458 (* 0.0272727 = 0.0942157 loss) | |
| I0525 02:13:13.250329 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.43612 (* 0.0272727 = 0.0937123 loss) | |
| I0525 02:13:13.250344 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.45908 (* 0.0272727 = 0.0943386 loss) | |
| I0525 02:13:13.250358 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.5491 (* 0.0272727 = 0.0967936 loss) | |
| I0525 02:13:13.250372 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.78663 (* 0.0272727 = 0.0487263 loss) | |
| I0525 02:13:13.250386 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.01118 (* 0.0272727 = 0.0275777 loss) | |
| I0525 02:13:13.250401 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.238145 (* 0.0272727 = 0.00649487 loss) | |
| I0525 02:13:13.250414 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.307087 (* 0.0272727 = 0.00837509 loss) | |
| I0525 02:13:13.250427 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.22737 (* 0.0272727 = 0.00620099 loss) | |
| I0525 02:13:13.250442 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.182438 (* 0.0272727 = 0.00497557 loss) | |
| I0525 02:13:13.250455 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.149378 (* 0.0272727 = 0.00407395 loss) | |
| I0525 02:13:13.250469 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.12995 (* 0.0272727 = 0.00354409 loss) | |
| I0525 02:13:13.250483 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.140215 (* 0.0272727 = 0.00382403 loss) | |
| I0525 02:13:13.250497 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0795217 (* 0.0272727 = 0.00216877 loss) | |
| I0525 02:13:13.250511 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0456341 (* 0.0272727 = 0.00124457 loss) | |
| I0525 02:13:13.250525 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0529744 (* 0.0272727 = 0.00144476 loss) | |
| I0525 02:13:13.250538 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0272023 (* 0.0272727 = 0.00074188 loss) | |
| I0525 02:13:13.250552 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0316644 (* 0.0272727 = 0.000863574 loss) | |
| I0525 02:13:13.250567 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0197323 (* 0.0272727 = 0.000538155 loss) | |
| I0525 02:13:13.250581 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0158742 (* 0.0272727 = 0.000432932 loss) | |
| I0525 02:13:13.250593 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0212766 | |
| I0525 02:13:13.250605 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 02:13:13.250617 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 02:13:13.250628 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 02:13:13.250640 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 02:13:13.250651 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 02:13:13.250663 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 02:13:13.250675 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 02:13:13.250687 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 02:13:13.250699 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 02:13:13.250711 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 02:13:13.250722 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 02:13:13.250733 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 02:13:13.250746 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 02:13:13.250757 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:13:13.250768 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:13:13.250779 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:13:13.250800 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:13:13.250814 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:13:13.250825 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:13:13.250834 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:13:13.250841 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:13:13.250854 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:13:13.250866 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.715909 | |
| I0525 02:13:13.250879 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.12766 | |
| I0525 02:13:13.250892 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.60341 (* 1 = 3.60341 loss) | |
| I0525 02:13:13.250906 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.14314 (* 1 = 1.14314 loss) | |
| I0525 02:13:13.250921 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.32619 (* 0.0909091 = 0.302381 loss) | |
| I0525 02:13:13.250936 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.48022 (* 0.0909091 = 0.316383 loss) | |
| I0525 02:13:13.250951 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.20842 (* 0.0909091 = 0.291674 loss) | |
| I0525 02:13:13.250964 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.62752 (* 0.0909091 = 0.329775 loss) | |
| I0525 02:13:13.250978 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.15356 (* 0.0909091 = 0.286687 loss) | |
| I0525 02:13:13.250991 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.56778 (* 0.0909091 = 0.324343 loss) | |
| I0525 02:13:13.251005 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.3118 (* 0.0909091 = 0.119254 loss) | |
| I0525 02:13:13.251019 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.724638 (* 0.0909091 = 0.0658762 loss) | |
| I0525 02:13:13.251034 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.151927 (* 0.0909091 = 0.0138115 loss) | |
| I0525 02:13:13.251047 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.11218 (* 0.0909091 = 0.0101982 loss) | |
| I0525 02:13:13.251061 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.089778 (* 0.0909091 = 0.00816164 loss) | |
| I0525 02:13:13.251075 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0723804 (* 0.0909091 = 0.00658003 loss) | |
| I0525 02:13:13.251090 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0299801 (* 0.0909091 = 0.00272546 loss) | |
| I0525 02:13:13.251103 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.022604 (* 0.0909091 = 0.00205491 loss) | |
| I0525 02:13:13.251117 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.017961 (* 0.0909091 = 0.00163282 loss) | |
| I0525 02:13:13.251132 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.010104 (* 0.0909091 = 0.000918546 loss) | |
| I0525 02:13:13.251144 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0100966 (* 0.0909091 = 0.000917876 loss) | |
| I0525 02:13:13.251158 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0063393 (* 0.0909091 = 0.0005763 loss) | |
| I0525 02:13:13.251173 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00228252 (* 0.0909091 = 0.000207501 loss) | |
| I0525 02:13:13.251186 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00215484 (* 0.0909091 = 0.000195894 loss) | |
| I0525 02:13:13.251200 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00235377 (* 0.0909091 = 0.000213979 loss) | |
| I0525 02:13:13.251214 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00204459 (* 0.0909091 = 0.000185872 loss) | |
| I0525 02:13:13.251225 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:13:13.251237 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:13:13.251248 5272 solver.cpp:245] Train net output #149: total_confidence = 0.000154486 | |
| I0525 02:13:13.251271 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000272474 | |
| I0525 02:13:13.251286 5272 sgd_solver.cpp:106] Iteration 9500, lr = 0.001 | |
| I0525 02:18:47.697013 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.7184 > 30) by scale factor 0.916916 | |
| I0525 02:19:37.720718 5272 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm20_iter_10000.caffemodel | |
| I0525 02:19:39.299392 5272 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm20_iter_10000.solverstate | |
| I0525 02:19:39.575157 5272 solver.cpp:338] Iteration 10000, Testing net (#0) | |
| I0525 02:20:37.407773 5272 solver.cpp:393] Test loss: 9.83334 | |
| I0525 02:20:37.407964 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0600036 | |
| I0525 02:20:37.407989 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.116 | |
| I0525 02:20:37.408004 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.075 | |
| I0525 02:20:37.408015 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.085 | |
| I0525 02:20:37.408028 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.155 | |
| I0525 02:20:37.408041 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.307 | |
| I0525 02:20:37.408052 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.47 | |
| I0525 02:20:37.408066 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.738 | |
| I0525 02:20:37.408077 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.92 | |
| I0525 02:20:37.408090 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.982 | |
| I0525 02:20:37.408102 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.994 | |
| I0525 02:20:37.408113 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0525 02:20:37.408125 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0525 02:20:37.408138 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0525 02:20:37.408149 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0525 02:20:37.408159 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0525 02:20:37.408171 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0525 02:20:37.408182 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0525 02:20:37.408193 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0525 02:20:37.408205 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0525 02:20:37.408216 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0525 02:20:37.408227 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0525 02:20:37.408238 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0525 02:20:37.408251 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.764183 | |
| I0525 02:20:37.408262 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.219971 | |
| I0525 02:20:37.408278 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.63939 (* 0.3 = 1.09182 loss) | |
| I0525 02:20:37.408294 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.946998 (* 0.3 = 0.284099 loss) | |
| I0525 02:20:37.408308 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 3.16998 (* 0.0272727 = 0.086454 loss) | |
| I0525 02:20:37.408323 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.35321 (* 0.0272727 = 0.091451 loss) | |
| I0525 02:20:37.408336 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.42513 (* 0.0272727 = 0.0934127 loss) | |
| I0525 02:20:37.408349 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.30432 (* 0.0272727 = 0.0901178 loss) | |
| I0525 02:20:37.408363 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 2.8424 (* 0.0272727 = 0.07752 loss) | |
| I0525 02:20:37.408377 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.37879 (* 0.0272727 = 0.0648761 loss) | |
| I0525 02:20:37.408391 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.40979 (* 0.0272727 = 0.0384489 loss) | |
| I0525 02:20:37.408404 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.49884 (* 0.0272727 = 0.0136047 loss) | |
| I0525 02:20:37.408418 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.139968 (* 0.0272727 = 0.0038173 loss) | |
| I0525 02:20:37.408432 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0824474 (* 0.0272727 = 0.00224857 loss) | |
| I0525 02:20:37.408447 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0398696 (* 0.0272727 = 0.00108735 loss) | |
| I0525 02:20:37.408460 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0353996 (* 0.0272727 = 0.000965444 loss) | |
| I0525 02:20:37.408474 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0237104 (* 0.0272727 = 0.000646648 loss) | |
| I0525 02:20:37.408502 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0189621 (* 0.0272727 = 0.000517149 loss) | |
| I0525 02:20:37.408517 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0155713 (* 0.0272727 = 0.000424672 loss) | |
| I0525 02:20:37.408531 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.00890219 (* 0.0272727 = 0.000242787 loss) | |
| I0525 02:20:37.408545 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0065179 (* 0.0272727 = 0.000177761 loss) | |
| I0525 02:20:37.408577 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00498743 (* 0.0272727 = 0.000136021 loss) | |
| I0525 02:20:37.408594 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00475685 (* 0.0272727 = 0.000129732 loss) | |
| I0525 02:20:37.408608 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00420073 (* 0.0272727 = 0.000114565 loss) | |
| I0525 02:20:37.408622 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00511386 (* 0.0272727 = 0.000139469 loss) | |
| I0525 02:20:37.408637 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00448325 (* 0.0272727 = 0.00012227 loss) | |
| I0525 02:20:37.408649 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0615175 | |
| I0525 02:20:37.408661 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.112 | |
| I0525 02:20:37.408674 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.08 | |
| I0525 02:20:37.408684 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.08 | |
| I0525 02:20:37.408697 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.16 | |
| I0525 02:20:37.408710 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.314 | |
| I0525 02:20:37.408720 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.471 | |
| I0525 02:20:37.408732 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.739 | |
| I0525 02:20:37.408743 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.919 | |
| I0525 02:20:37.408754 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.982 | |
| I0525 02:20:37.408766 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.994 | |
| I0525 02:20:37.408777 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0525 02:20:37.408788 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0525 02:20:37.408800 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0525 02:20:37.408812 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0525 02:20:37.408823 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0525 02:20:37.408833 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0525 02:20:37.408844 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0525 02:20:37.408855 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0525 02:20:37.408866 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0525 02:20:37.408880 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0525 02:20:37.408892 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0525 02:20:37.408903 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0525 02:20:37.408915 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.764001 | |
| I0525 02:20:37.408926 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.21614 | |
| I0525 02:20:37.408939 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.61726 (* 0.3 = 1.08518 loss) | |
| I0525 02:20:37.408953 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.94186 (* 0.3 = 0.282558 loss) | |
| I0525 02:20:37.408969 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 3.17361 (* 0.0272727 = 0.0865531 loss) | |
| I0525 02:20:37.408983 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.35399 (* 0.0272727 = 0.0914724 loss) | |
| I0525 02:20:37.408998 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.42504 (* 0.0272727 = 0.0934102 loss) | |
| I0525 02:20:37.409023 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 3.32007 (* 0.0272727 = 0.0905474 loss) | |
| I0525 02:20:37.409037 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 2.86623 (* 0.0272727 = 0.0781699 loss) | |
| I0525 02:20:37.409051 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.3996 (* 0.0272727 = 0.0654436 loss) | |
| I0525 02:20:37.409065 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.41522 (* 0.0272727 = 0.0385969 loss) | |
| I0525 02:20:37.409078 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.514244 (* 0.0272727 = 0.0140248 loss) | |
| I0525 02:20:37.409092 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.150824 (* 0.0272727 = 0.00411339 loss) | |
| I0525 02:20:37.409106 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0902685 (* 0.0272727 = 0.00246187 loss) | |
| I0525 02:20:37.409133 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.043592 (* 0.0272727 = 0.00118887 loss) | |
| I0525 02:20:37.409149 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0331394 (* 0.0272727 = 0.000903802 loss) | |
| I0525 02:20:37.409163 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0255287 (* 0.0272727 = 0.000696237 loss) | |
| I0525 02:20:37.409178 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.0193416 (* 0.0272727 = 0.000527498 loss) | |
| I0525 02:20:37.409190 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.0148862 (* 0.0272727 = 0.000405987 loss) | |
| I0525 02:20:37.409204 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00954894 (* 0.0272727 = 0.000260426 loss) | |
| I0525 02:20:37.409219 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.005174 (* 0.0272727 = 0.000141109 loss) | |
| I0525 02:20:37.409232 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00519823 (* 0.0272727 = 0.00014177 loss) | |
| I0525 02:20:37.409245 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.00417005 (* 0.0272727 = 0.000113729 loss) | |
| I0525 02:20:37.409260 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00436042 (* 0.0272727 = 0.000118921 loss) | |
| I0525 02:20:37.409272 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00414404 (* 0.0272727 = 0.000113019 loss) | |
| I0525 02:20:37.409286 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00373517 (* 0.0272727 = 0.000101868 loss) | |
| I0525 02:20:37.409298 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0776722 | |
| I0525 02:20:37.409307 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.101 | |
| I0525 02:20:37.409314 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.091 | |
| I0525 02:20:37.409322 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.082 | |
| I0525 02:20:37.409334 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.138 | |
| I0525 02:20:37.409345 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.306 | |
| I0525 02:20:37.409358 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.457 | |
| I0525 02:20:37.409368 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.73 | |
| I0525 02:20:37.409380 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.91 | |
| I0525 02:20:37.409391 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.979 | |
| I0525 02:20:37.409404 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.991 | |
| I0525 02:20:37.409415 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0525 02:20:37.409425 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0525 02:20:37.409436 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0525 02:20:37.409447 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0525 02:20:37.409458 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0525 02:20:37.409469 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0525 02:20:37.409481 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0525 02:20:37.409503 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0525 02:20:37.409515 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0525 02:20:37.409528 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0525 02:20:37.409538 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0525 02:20:37.409549 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0525 02:20:37.409560 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.755046 | |
| I0525 02:20:37.409572 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.233118 | |
| I0525 02:20:37.409585 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.20213 (* 1 = 3.20213 loss) | |
| I0525 02:20:37.409600 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.938449 (* 1 = 0.938449 loss) | |
| I0525 02:20:37.409612 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 2.97184 (* 0.0909091 = 0.270167 loss) | |
| I0525 02:20:37.409626 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 3.17801 (* 0.0909091 = 0.28891 loss) | |
| I0525 02:20:37.409639 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.25949 (* 0.0909091 = 0.296317 loss) | |
| I0525 02:20:37.409653 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 3.15739 (* 0.0909091 = 0.287035 loss) | |
| I0525 02:20:37.409667 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.73003 (* 0.0909091 = 0.248185 loss) | |
| I0525 02:20:37.409680 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.3041 (* 0.0909091 = 0.209464 loss) | |
| I0525 02:20:37.409693 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.34969 (* 0.0909091 = 0.122699 loss) | |
| I0525 02:20:37.409706 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.555437 (* 0.0909091 = 0.0504943 loss) | |
| I0525 02:20:37.409719 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.157872 (* 0.0909091 = 0.014352 loss) | |
| I0525 02:20:37.409734 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0930276 (* 0.0909091 = 0.00845705 loss) | |
| I0525 02:20:37.409747 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0497495 (* 0.0909091 = 0.00452268 loss) | |
| I0525 02:20:37.409760 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0397187 (* 0.0909091 = 0.00361079 loss) | |
| I0525 02:20:37.409775 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0305851 (* 0.0909091 = 0.00278047 loss) | |
| I0525 02:20:37.409787 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.0235545 (* 0.0909091 = 0.00214132 loss) | |
| I0525 02:20:37.409801 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.0163944 (* 0.0909091 = 0.0014904 loss) | |
| I0525 02:20:37.409814 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00962252 (* 0.0909091 = 0.000874775 loss) | |
| I0525 02:20:37.409828 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00544977 (* 0.0909091 = 0.000495433 loss) | |
| I0525 02:20:37.409842 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00328961 (* 0.0909091 = 0.000299055 loss) | |
| I0525 02:20:37.409855 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00246117 (* 0.0909091 = 0.000223742 loss) | |
| I0525 02:20:37.409868 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00178583 (* 0.0909091 = 0.000162348 loss) | |
| I0525 02:20:37.409883 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00155015 (* 0.0909091 = 0.000140923 loss) | |
| I0525 02:20:37.409895 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00135665 (* 0.0909091 = 0.000123332 loss) | |
| I0525 02:20:37.409907 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 02:20:37.409919 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:20:37.409932 5272 solver.cpp:406] Test net output #149: total_confidence = 2.20222e-05 | |
| I0525 02:20:37.409945 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 6.82075e-05 | |
| I0525 02:20:37.409968 5272 solver.cpp:338] Iteration 10000, Testing net (#1) | |
| I0525 02:21:35.310981 5272 solver.cpp:393] Test loss: 10.465 | |
| I0525 02:21:35.311115 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0578559 | |
| I0525 02:21:35.311136 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.109 | |
| I0525 02:21:35.311151 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.112 | |
| I0525 02:21:35.311162 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.084 | |
| I0525 02:21:35.311175 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.168 | |
| I0525 02:21:35.311187 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.32 | |
| I0525 02:21:35.311200 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.441 | |
| I0525 02:21:35.311213 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.656 | |
| I0525 02:21:35.311224 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.825 | |
| I0525 02:21:35.311238 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.887 | |
| I0525 02:21:35.311249 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.902 | |
| I0525 02:21:35.311261 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.926 | |
| I0525 02:21:35.311274 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.943 | |
| I0525 02:21:35.311285 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.952 | |
| I0525 02:21:35.311296 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.963 | |
| I0525 02:21:35.311308 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.966 | |
| I0525 02:21:35.311321 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.982 | |
| I0525 02:21:35.311332 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.992 | |
| I0525 02:21:35.311343 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.993 | |
| I0525 02:21:35.311355 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.996 | |
| I0525 02:21:35.311367 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.999 | |
| I0525 02:21:35.311378 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.999 | |
| I0525 02:21:35.311390 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.999 | |
| I0525 02:21:35.311401 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.73191 | |
| I0525 02:21:35.311414 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.219995 | |
| I0525 02:21:35.311429 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.68745 (* 0.3 = 1.10624 loss) | |
| I0525 02:21:35.311444 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.09099 (* 0.3 = 0.327298 loss) | |
| I0525 02:21:35.311458 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 3.24691 (* 0.0272727 = 0.0885522 loss) | |
| I0525 02:21:35.311472 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.3158 (* 0.0272727 = 0.0904308 loss) | |
| I0525 02:21:35.311486 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.43561 (* 0.0272727 = 0.0936985 loss) | |
| I0525 02:21:35.311499 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.24865 (* 0.0272727 = 0.0885995 loss) | |
| I0525 02:21:35.311513 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 2.82589 (* 0.0272727 = 0.0770696 loss) | |
| I0525 02:21:35.311527 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.49535 (* 0.0272727 = 0.0680549 loss) | |
| I0525 02:21:35.311540 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.66249 (* 0.0272727 = 0.0453406 loss) | |
| I0525 02:21:35.311554 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.874205 (* 0.0272727 = 0.0238419 loss) | |
| I0525 02:21:35.311568 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.544059 (* 0.0272727 = 0.014838 loss) | |
| I0525 02:21:35.311581 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.47089 (* 0.0272727 = 0.0128425 loss) | |
| I0525 02:21:35.311595 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.371633 (* 0.0272727 = 0.0101354 loss) | |
| I0525 02:21:35.311609 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.297776 (* 0.0272727 = 0.00812116 loss) | |
| I0525 02:21:35.311642 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.257627 (* 0.0272727 = 0.00702618 loss) | |
| I0525 02:21:35.311657 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.210675 (* 0.0272727 = 0.00574567 loss) | |
| I0525 02:21:35.311671 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.192944 (* 0.0272727 = 0.00526211 loss) | |
| I0525 02:21:35.311686 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.121605 (* 0.0272727 = 0.0033165 loss) | |
| I0525 02:21:35.311699 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0683839 (* 0.0272727 = 0.00186502 loss) | |
| I0525 02:21:35.311713 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0591718 (* 0.0272727 = 0.00161378 loss) | |
| I0525 02:21:35.311728 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.038888 (* 0.0272727 = 0.00106058 loss) | |
| I0525 02:21:35.311741 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0154692 (* 0.0272727 = 0.000421886 loss) | |
| I0525 02:21:35.311755 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0160951 (* 0.0272727 = 0.000438958 loss) | |
| I0525 02:21:35.311769 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0141512 (* 0.0272727 = 0.000385941 loss) | |
| I0525 02:21:35.311780 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.062623 | |
| I0525 02:21:35.311792 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.119 | |
| I0525 02:21:35.311805 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.104 | |
| I0525 02:21:35.311815 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.081 | |
| I0525 02:21:35.311827 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.16 | |
| I0525 02:21:35.311839 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.326 | |
| I0525 02:21:35.311851 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.444 | |
| I0525 02:21:35.311862 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.654 | |
| I0525 02:21:35.311877 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.824 | |
| I0525 02:21:35.311889 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.887 | |
| I0525 02:21:35.311902 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.902 | |
| I0525 02:21:35.311913 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.926 | |
| I0525 02:21:35.311925 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.943 | |
| I0525 02:21:35.311936 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.952 | |
| I0525 02:21:35.311947 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.963 | |
| I0525 02:21:35.311959 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.966 | |
| I0525 02:21:35.311970 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.982 | |
| I0525 02:21:35.311981 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.992 | |
| I0525 02:21:35.311992 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.993 | |
| I0525 02:21:35.312005 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.996 | |
| I0525 02:21:35.312016 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.999 | |
| I0525 02:21:35.312027 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.999 | |
| I0525 02:21:35.312038 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.999 | |
| I0525 02:21:35.312050 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.733455 | |
| I0525 02:21:35.312062 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.211085 | |
| I0525 02:21:35.312075 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.63767 (* 0.3 = 1.0913 loss) | |
| I0525 02:21:35.312089 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.07434 (* 0.3 = 0.322303 loss) | |
| I0525 02:21:35.312103 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 3.25445 (* 0.0272727 = 0.0887578 loss) | |
| I0525 02:21:35.312116 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.31663 (* 0.0272727 = 0.0904535 loss) | |
| I0525 02:21:35.312144 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.42929 (* 0.0272727 = 0.093526 loss) | |
| I0525 02:21:35.312158 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 3.25517 (* 0.0272727 = 0.0887773 loss) | |
| I0525 02:21:35.312172 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 2.8289 (* 0.0272727 = 0.0771517 loss) | |
| I0525 02:21:35.312186 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.50525 (* 0.0272727 = 0.068325 loss) | |
| I0525 02:21:35.312199 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.66849 (* 0.0272727 = 0.0455044 loss) | |
| I0525 02:21:35.312212 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.881974 (* 0.0272727 = 0.0240538 loss) | |
| I0525 02:21:35.312225 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.562683 (* 0.0272727 = 0.0153459 loss) | |
| I0525 02:21:35.312239 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.471991 (* 0.0272727 = 0.0128725 loss) | |
| I0525 02:21:35.312253 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.368717 (* 0.0272727 = 0.0100559 loss) | |
| I0525 02:21:35.312266 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.302846 (* 0.0272727 = 0.00825942 loss) | |
| I0525 02:21:35.312280 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.261626 (* 0.0272727 = 0.00713525 loss) | |
| I0525 02:21:35.312304 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.210753 (* 0.0272727 = 0.00574782 loss) | |
| I0525 02:21:35.312327 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.199207 (* 0.0272727 = 0.00543293 loss) | |
| I0525 02:21:35.312342 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.125161 (* 0.0272727 = 0.00341348 loss) | |
| I0525 02:21:35.312356 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0668593 (* 0.0272727 = 0.00182343 loss) | |
| I0525 02:21:35.312369 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.06017 (* 0.0272727 = 0.001641 loss) | |
| I0525 02:21:35.312383 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0372799 (* 0.0272727 = 0.00101672 loss) | |
| I0525 02:21:35.312397 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0137686 (* 0.0272727 = 0.000375508 loss) | |
| I0525 02:21:35.312412 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0139979 (* 0.0272727 = 0.000381762 loss) | |
| I0525 02:21:35.312424 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0144818 (* 0.0272727 = 0.000394959 loss) | |
| I0525 02:21:35.312436 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0825973 | |
| I0525 02:21:35.312448 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.119 | |
| I0525 02:21:35.312459 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.101 | |
| I0525 02:21:35.312471 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.09 | |
| I0525 02:21:35.312482 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.154 | |
| I0525 02:21:35.312494 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.319 | |
| I0525 02:21:35.312505 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.442 | |
| I0525 02:21:35.312513 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.65 | |
| I0525 02:21:35.312520 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.823 | |
| I0525 02:21:35.312532 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.886 | |
| I0525 02:21:35.312544 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.898 | |
| I0525 02:21:35.312556 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.925 | |
| I0525 02:21:35.312566 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.943 | |
| I0525 02:21:35.312578 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.952 | |
| I0525 02:21:35.312589 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.963 | |
| I0525 02:21:35.312600 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.966 | |
| I0525 02:21:35.312613 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.982 | |
| I0525 02:21:35.312634 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.992 | |
| I0525 02:21:35.312647 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.993 | |
| I0525 02:21:35.312659 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.996 | |
| I0525 02:21:35.312670 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.999 | |
| I0525 02:21:35.312681 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.999 | |
| I0525 02:21:35.312693 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.999 | |
| I0525 02:21:35.312705 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.728682 | |
| I0525 02:21:35.312716 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.224672 | |
| I0525 02:21:35.312731 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.21502 (* 1 = 3.21502 loss) | |
| I0525 02:21:35.312755 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 1.0456 (* 1 = 1.0456 loss) | |
| I0525 02:21:35.312777 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 3.06666 (* 0.0909091 = 0.278787 loss) | |
| I0525 02:21:35.312793 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 3.14282 (* 0.0909091 = 0.285711 loss) | |
| I0525 02:21:35.312805 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.25603 (* 0.0909091 = 0.296003 loss) | |
| I0525 02:21:35.312819 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 3.11557 (* 0.0909091 = 0.283234 loss) | |
| I0525 02:21:35.312832 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.69726 (* 0.0909091 = 0.245205 loss) | |
| I0525 02:21:35.312846 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.36823 (* 0.0909091 = 0.215294 loss) | |
| I0525 02:21:35.312860 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.57995 (* 0.0909091 = 0.143631 loss) | |
| I0525 02:21:35.312872 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.88992 (* 0.0909091 = 0.0809018 loss) | |
| I0525 02:21:35.312885 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.522865 (* 0.0909091 = 0.0475332 loss) | |
| I0525 02:21:35.312899 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.44528 (* 0.0909091 = 0.04048 loss) | |
| I0525 02:21:35.312913 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.347268 (* 0.0909091 = 0.0315698 loss) | |
| I0525 02:21:35.312929 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.283612 (* 0.0909091 = 0.025783 loss) | |
| I0525 02:21:35.312943 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.243715 (* 0.0909091 = 0.0221559 loss) | |
| I0525 02:21:35.312958 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.204154 (* 0.0909091 = 0.0185594 loss) | |
| I0525 02:21:35.312970 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.177459 (* 0.0909091 = 0.0161326 loss) | |
| I0525 02:21:35.312984 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.108902 (* 0.0909091 = 0.00990016 loss) | |
| I0525 02:21:35.312997 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0565278 (* 0.0909091 = 0.00513889 loss) | |
| I0525 02:21:35.313011 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0571759 (* 0.0909091 = 0.00519781 loss) | |
| I0525 02:21:35.313024 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0361376 (* 0.0909091 = 0.00328523 loss) | |
| I0525 02:21:35.313038 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0121723 (* 0.0909091 = 0.00110657 loss) | |
| I0525 02:21:35.313051 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0139018 (* 0.0909091 = 0.0012638 loss) | |
| I0525 02:21:35.313066 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.0136048 (* 0.0909091 = 0.0012368 loss) | |
| I0525 02:21:35.313076 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 02:21:35.313088 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:21:35.313098 5272 solver.cpp:406] Test net output #149: total_confidence = 2.49571e-05 | |
| I0525 02:21:35.313136 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 8.22276e-05 | |
| I0525 02:21:35.671535 5272 solver.cpp:229] Iteration 10000, loss = 10.4865 | |
| I0525 02:21:35.671607 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0681818 | |
| I0525 02:21:35.671625 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 02:21:35.671639 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 02:21:35.671653 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 02:21:35.671665 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 02:21:35.671677 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 02:21:35.671690 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 02:21:35.671703 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 02:21:35.671717 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 02:21:35.671730 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 02:21:35.671743 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 02:21:35.671756 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 02:21:35.671767 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 02:21:35.671782 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 02:21:35.671795 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:21:35.671807 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:21:35.671819 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:21:35.671831 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:21:35.671844 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:21:35.671855 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:21:35.671867 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:21:35.671880 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:21:35.671891 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:21:35.671905 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0525 02:21:35.671916 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.272727 | |
| I0525 02:21:35.671933 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.22828 (* 0.3 = 0.968483 loss) | |
| I0525 02:21:35.671947 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.01134 (* 0.3 = 0.303401 loss) | |
| I0525 02:21:35.671962 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.62114 (* 0.0272727 = 0.0987584 loss) | |
| I0525 02:21:35.671977 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.42471 (* 0.0272727 = 0.0934013 loss) | |
| I0525 02:21:35.671990 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.56778 (* 0.0272727 = 0.097303 loss) | |
| I0525 02:21:35.672004 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.21064 (* 0.0272727 = 0.087563 loss) | |
| I0525 02:21:35.672019 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.72688 (* 0.0272727 = 0.0743696 loss) | |
| I0525 02:21:35.672032 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.99881 (* 0.0272727 = 0.054513 loss) | |
| I0525 02:21:35.672046 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.34254 (* 0.0272727 = 0.0366147 loss) | |
| I0525 02:21:35.672061 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.12287 (* 0.0272727 = 0.0306238 loss) | |
| I0525 02:21:35.672075 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.107967 (* 0.0272727 = 0.00294454 loss) | |
| I0525 02:21:35.672089 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0713379 (* 0.0272727 = 0.00194558 loss) | |
| I0525 02:21:35.672104 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0545957 (* 0.0272727 = 0.00148897 loss) | |
| I0525 02:21:35.672153 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0371587 (* 0.0272727 = 0.00101342 loss) | |
| I0525 02:21:35.672168 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0499097 (* 0.0272727 = 0.00136117 loss) | |
| I0525 02:21:35.672183 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0293028 (* 0.0272727 = 0.000799168 loss) | |
| I0525 02:21:35.672196 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0119811 (* 0.0272727 = 0.000326756 loss) | |
| I0525 02:21:35.672210 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0178584 (* 0.0272727 = 0.000487048 loss) | |
| I0525 02:21:35.672224 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0074372 (* 0.0272727 = 0.000202833 loss) | |
| I0525 02:21:35.672238 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00500947 (* 0.0272727 = 0.000136622 loss) | |
| I0525 02:21:35.672253 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00422975 (* 0.0272727 = 0.000115357 loss) | |
| I0525 02:21:35.672268 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00396952 (* 0.0272727 = 0.00010826 loss) | |
| I0525 02:21:35.672283 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00366643 (* 0.0272727 = 9.99936e-05 loss) | |
| I0525 02:21:35.672297 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00524919 (* 0.0272727 = 0.00014316 loss) | |
| I0525 02:21:35.672310 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0681818 | |
| I0525 02:21:35.672322 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 02:21:35.672334 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 02:21:35.672346 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 02:21:35.672358 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 02:21:35.672369 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 02:21:35.672381 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 02:21:35.672394 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 02:21:35.672405 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 02:21:35.672417 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 02:21:35.672430 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 02:21:35.672444 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 02:21:35.672456 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 02:21:35.672468 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 02:21:35.672480 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:21:35.672492 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:21:35.672503 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:21:35.672515 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:21:35.672526 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:21:35.672538 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:21:35.672549 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:21:35.672561 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:21:35.672574 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:21:35.672585 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.744318 | |
| I0525 02:21:35.672596 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.318182 | |
| I0525 02:21:35.672610 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.19535 (* 0.3 = 0.958606 loss) | |
| I0525 02:21:35.672624 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.04085 (* 0.3 = 0.312256 loss) | |
| I0525 02:21:35.672638 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.95244 (* 0.0272727 = 0.107794 loss) | |
| I0525 02:21:35.672663 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.0055 (* 0.0272727 = 0.0819681 loss) | |
| I0525 02:21:35.672678 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.66933 (* 0.0272727 = 0.100073 loss) | |
| I0525 02:21:35.672693 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.07422 (* 0.0272727 = 0.0838424 loss) | |
| I0525 02:21:35.672706 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.02989 (* 0.0272727 = 0.0826333 loss) | |
| I0525 02:21:35.672719 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.14249 (* 0.0272727 = 0.0584315 loss) | |
| I0525 02:21:35.672734 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.31733 (* 0.0272727 = 0.0359272 loss) | |
| I0525 02:21:35.672746 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.00943 (* 0.0272727 = 0.02753 loss) | |
| I0525 02:21:35.672760 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0678745 (* 0.0272727 = 0.00185112 loss) | |
| I0525 02:21:35.672775 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0445991 (* 0.0272727 = 0.00121634 loss) | |
| I0525 02:21:35.672788 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0424249 (* 0.0272727 = 0.00115704 loss) | |
| I0525 02:21:35.672799 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.038086 (* 0.0272727 = 0.00103871 loss) | |
| I0525 02:21:35.672809 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0265165 (* 0.0272727 = 0.000723177 loss) | |
| I0525 02:21:35.672827 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0325395 (* 0.0272727 = 0.00088744 loss) | |
| I0525 02:21:35.672840 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0165017 (* 0.0272727 = 0.000450045 loss) | |
| I0525 02:21:35.672855 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0179028 (* 0.0272727 = 0.000488257 loss) | |
| I0525 02:21:35.672869 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00836521 (* 0.0272727 = 0.000228142 loss) | |
| I0525 02:21:35.672883 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00924377 (* 0.0272727 = 0.000252103 loss) | |
| I0525 02:21:35.672897 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00869589 (* 0.0272727 = 0.000237161 loss) | |
| I0525 02:21:35.672911 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00726901 (* 0.0272727 = 0.000198246 loss) | |
| I0525 02:21:35.672925 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0099611 (* 0.0272727 = 0.000271666 loss) | |
| I0525 02:21:35.672938 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00892544 (* 0.0272727 = 0.000243421 loss) | |
| I0525 02:21:35.672951 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0454545 | |
| I0525 02:21:35.672963 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 02:21:35.672976 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 02:21:35.672987 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0525 02:21:35.672999 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 02:21:35.673012 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 02:21:35.673023 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 02:21:35.673035 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 02:21:35.673048 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 02:21:35.673058 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 02:21:35.673070 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 02:21:35.673082 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 02:21:35.673094 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 02:21:35.673105 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 02:21:35.673128 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:21:35.673154 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:21:35.673168 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:21:35.673180 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:21:35.673192 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:21:35.673203 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:21:35.673215 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:21:35.673226 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:21:35.673238 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:21:35.673250 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.727273 | |
| I0525 02:21:35.673262 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.204545 | |
| I0525 02:21:35.673276 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.27141 (* 1 = 3.27141 loss) | |
| I0525 02:21:35.673290 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.05304 (* 1 = 1.05304 loss) | |
| I0525 02:21:35.673305 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.41528 (* 0.0909091 = 0.31048 loss) | |
| I0525 02:21:35.673318 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.12796 (* 0.0909091 = 0.28436 loss) | |
| I0525 02:21:35.673331 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.84016 (* 0.0909091 = 0.349105 loss) | |
| I0525 02:21:35.673346 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.85887 (* 0.0909091 = 0.259898 loss) | |
| I0525 02:21:35.673359 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.57313 (* 0.0909091 = 0.233921 loss) | |
| I0525 02:21:35.673373 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.79931 (* 0.0909091 = 0.163574 loss) | |
| I0525 02:21:35.673387 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.43675 (* 0.0909091 = 0.130614 loss) | |
| I0525 02:21:35.673400 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.880615 (* 0.0909091 = 0.0800559 loss) | |
| I0525 02:21:35.673413 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.109578 (* 0.0909091 = 0.00996167 loss) | |
| I0525 02:21:35.673427 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0728514 (* 0.0909091 = 0.00662285 loss) | |
| I0525 02:21:35.673441 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0556163 (* 0.0909091 = 0.00505603 loss) | |
| I0525 02:21:35.673455 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0410828 (* 0.0909091 = 0.0037348 loss) | |
| I0525 02:21:35.673470 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0222269 (* 0.0909091 = 0.00202063 loss) | |
| I0525 02:21:35.673483 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0348786 (* 0.0909091 = 0.00317078 loss) | |
| I0525 02:21:35.673502 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0149971 (* 0.0909091 = 0.00136338 loss) | |
| I0525 02:21:35.673516 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0110018 (* 0.0909091 = 0.00100017 loss) | |
| I0525 02:21:35.673530 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00713083 (* 0.0909091 = 0.000648258 loss) | |
| I0525 02:21:35.673547 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00472702 (* 0.0909091 = 0.000429729 loss) | |
| I0525 02:21:35.673560 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00420165 (* 0.0909091 = 0.000381968 loss) | |
| I0525 02:21:35.673574 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00305481 (* 0.0909091 = 0.00027771 loss) | |
| I0525 02:21:35.673588 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00313771 (* 0.0909091 = 0.000285246 loss) | |
| I0525 02:21:35.673602 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00216713 (* 0.0909091 = 0.000197012 loss) | |
| I0525 02:21:35.673614 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:21:35.673636 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:21:35.673650 5272 solver.cpp:245] Train net output #149: total_confidence = 1.41469e-05 | |
| I0525 02:21:35.673661 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000203692 | |
| I0525 02:21:35.673676 5272 sgd_solver.cpp:106] Iteration 10000, lr = 0.001 | |
| I0525 02:24:46.944454 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.6155 > 30) by scale factor 0.819325 | |
| I0525 02:26:56.324043 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.5703 > 30) by scale factor 0.867798 | |
| I0525 02:28:00.645170 5272 solver.cpp:229] Iteration 10500, loss = 10.5161 | |
| I0525 02:28:00.645267 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.113636 | |
| I0525 02:28:00.645287 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 02:28:00.645300 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 02:28:00.645313 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 02:28:00.645325 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 02:28:00.645337 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 02:28:00.645350 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 02:28:00.645362 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 02:28:00.645375 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 02:28:00.645387 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 02:28:00.645401 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 02:28:00.645412 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 02:28:00.645424 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 02:28:00.645437 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 02:28:00.645448 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:28:00.645459 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:28:00.645472 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:28:00.645483 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:28:00.645495 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:28:00.645506 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:28:00.645519 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:28:00.645530 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:28:00.645542 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:28:00.645555 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.755682 | |
| I0525 02:28:00.645566 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.227273 | |
| I0525 02:28:00.645583 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.93714 (* 0.3 = 1.18114 loss) | |
| I0525 02:28:00.645597 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.19573 (* 0.3 = 0.358719 loss) | |
| I0525 02:28:00.645612 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 4.19738 (* 0.0272727 = 0.114474 loss) | |
| I0525 02:28:00.645627 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 4.30166 (* 0.0272727 = 0.117318 loss) | |
| I0525 02:28:00.645640 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.65508 (* 0.0272727 = 0.126957 loss) | |
| I0525 02:28:00.645654 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.8046 (* 0.0272727 = 0.103762 loss) | |
| I0525 02:28:00.645668 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 4.07525 (* 0.0272727 = 0.111143 loss) | |
| I0525 02:28:00.645683 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.25911 (* 0.0272727 = 0.0888847 loss) | |
| I0525 02:28:00.645696 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.09576 (* 0.0272727 = 0.0298844 loss) | |
| I0525 02:28:00.645710 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.109662 (* 0.0272727 = 0.00299079 loss) | |
| I0525 02:28:00.645725 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0150507 (* 0.0272727 = 0.000410473 loss) | |
| I0525 02:28:00.645745 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.020037 (* 0.0272727 = 0.000546464 loss) | |
| I0525 02:28:00.645761 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0149847 (* 0.0272727 = 0.000408674 loss) | |
| I0525 02:28:00.645774 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0141346 (* 0.0272727 = 0.00038549 loss) | |
| I0525 02:28:00.645788 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0136142 (* 0.0272727 = 0.000371296 loss) | |
| I0525 02:28:00.645823 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0160565 (* 0.0272727 = 0.000437904 loss) | |
| I0525 02:28:00.645839 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.010644 (* 0.0272727 = 0.000290291 loss) | |
| I0525 02:28:00.645853 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00852614 (* 0.0272727 = 0.000232531 loss) | |
| I0525 02:28:00.645869 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00958703 (* 0.0272727 = 0.000261464 loss) | |
| I0525 02:28:00.645882 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00604765 (* 0.0272727 = 0.000164936 loss) | |
| I0525 02:28:00.645896 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00766853 (* 0.0272727 = 0.000209142 loss) | |
| I0525 02:28:00.645911 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00471937 (* 0.0272727 = 0.00012871 loss) | |
| I0525 02:28:00.645925 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00935873 (* 0.0272727 = 0.000255238 loss) | |
| I0525 02:28:00.645938 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0117441 (* 0.0272727 = 0.000320292 loss) | |
| I0525 02:28:00.645951 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0227273 | |
| I0525 02:28:00.645963 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 02:28:00.645975 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 02:28:00.645987 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 02:28:00.645998 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 02:28:00.646010 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0525 02:28:00.646021 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 02:28:00.646034 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 02:28:00.646045 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 02:28:00.646056 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 02:28:00.646069 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 02:28:00.646080 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 02:28:00.646091 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 02:28:00.646102 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 02:28:00.646114 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:28:00.646126 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:28:00.646137 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:28:00.646148 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:28:00.646159 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:28:00.646172 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:28:00.646183 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:28:00.646194 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:28:00.646205 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:28:00.646217 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.738636 | |
| I0525 02:28:00.646229 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.181818 | |
| I0525 02:28:00.646244 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.91437 (* 0.3 = 1.17431 loss) | |
| I0525 02:28:00.646257 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.15735 (* 0.3 = 0.347204 loss) | |
| I0525 02:28:00.646270 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.16786 (* 0.0272727 = 0.113669 loss) | |
| I0525 02:28:00.646284 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.08381 (* 0.0272727 = 0.111377 loss) | |
| I0525 02:28:00.646297 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.65205 (* 0.0272727 = 0.126874 loss) | |
| I0525 02:28:00.646322 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.50792 (* 0.0272727 = 0.0956707 loss) | |
| I0525 02:28:00.646337 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.84236 (* 0.0272727 = 0.104792 loss) | |
| I0525 02:28:00.646352 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.9908 (* 0.0272727 = 0.0815671 loss) | |
| I0525 02:28:00.646365 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 0.923318 (* 0.0272727 = 0.0251814 loss) | |
| I0525 02:28:00.646379 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0650213 (* 0.0272727 = 0.00177331 loss) | |
| I0525 02:28:00.646394 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0178501 (* 0.0272727 = 0.000486821 loss) | |
| I0525 02:28:00.646409 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0148369 (* 0.0272727 = 0.000404642 loss) | |
| I0525 02:28:00.646422 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0104599 (* 0.0272727 = 0.000285269 loss) | |
| I0525 02:28:00.646436 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0070252 (* 0.0272727 = 0.000191596 loss) | |
| I0525 02:28:00.646450 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0156712 (* 0.0272727 = 0.000427395 loss) | |
| I0525 02:28:00.646464 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00604206 (* 0.0272727 = 0.000164783 loss) | |
| I0525 02:28:00.646478 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00978446 (* 0.0272727 = 0.000266849 loss) | |
| I0525 02:28:00.646492 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00681665 (* 0.0272727 = 0.000185909 loss) | |
| I0525 02:28:00.646507 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00238634 (* 0.0272727 = 6.5082e-05 loss) | |
| I0525 02:28:00.646520 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00880411 (* 0.0272727 = 0.000240112 loss) | |
| I0525 02:28:00.646534 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00385024 (* 0.0272727 = 0.000105007 loss) | |
| I0525 02:28:00.646548 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00453315 (* 0.0272727 = 0.000123632 loss) | |
| I0525 02:28:00.646561 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00390887 (* 0.0272727 = 0.000106605 loss) | |
| I0525 02:28:00.646575 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00390801 (* 0.0272727 = 0.000106582 loss) | |
| I0525 02:28:00.646587 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0681818 | |
| I0525 02:28:00.646600 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 02:28:00.646611 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0525 02:28:00.646623 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 02:28:00.646636 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 02:28:00.646647 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 02:28:00.646659 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 02:28:00.646672 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 02:28:00.646683 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 02:28:00.646694 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 02:28:00.646705 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 02:28:00.646718 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 02:28:00.646729 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 02:28:00.646740 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 02:28:00.646751 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:28:00.646764 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:28:00.646775 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:28:00.646801 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:28:00.646816 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:28:00.646827 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:28:00.646839 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:28:00.646850 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:28:00.646862 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:28:00.646873 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.75 | |
| I0525 02:28:00.646886 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.227273 | |
| I0525 02:28:00.646899 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.70604 (* 1 = 3.70604 loss) | |
| I0525 02:28:00.646914 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.11588 (* 1 = 1.11588 loss) | |
| I0525 02:28:00.646927 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 4.00319 (* 0.0909091 = 0.363926 loss) | |
| I0525 02:28:00.646941 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.61836 (* 0.0909091 = 0.328941 loss) | |
| I0525 02:28:00.646955 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 4.3406 (* 0.0909091 = 0.3946 loss) | |
| I0525 02:28:00.646970 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.67213 (* 0.0909091 = 0.33383 loss) | |
| I0525 02:28:00.646982 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 4.03098 (* 0.0909091 = 0.366453 loss) | |
| I0525 02:28:00.646997 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.87937 (* 0.0909091 = 0.261761 loss) | |
| I0525 02:28:00.647011 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.748941 (* 0.0909091 = 0.0680856 loss) | |
| I0525 02:28:00.647025 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0794521 (* 0.0909091 = 0.00722292 loss) | |
| I0525 02:28:00.647039 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0244644 (* 0.0909091 = 0.00222403 loss) | |
| I0525 02:28:00.647053 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0144105 (* 0.0909091 = 0.00131005 loss) | |
| I0525 02:28:00.647068 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0109486 (* 0.0909091 = 0.00099533 loss) | |
| I0525 02:28:00.647080 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00609529 (* 0.0909091 = 0.000554118 loss) | |
| I0525 02:28:00.647095 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00528298 (* 0.0909091 = 0.000480271 loss) | |
| I0525 02:28:00.647109 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00484999 (* 0.0909091 = 0.000440909 loss) | |
| I0525 02:28:00.647119 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00344551 (* 0.0909091 = 0.000313228 loss) | |
| I0525 02:28:00.647130 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0030429 (* 0.0909091 = 0.000276627 loss) | |
| I0525 02:28:00.647145 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00524567 (* 0.0909091 = 0.000476879 loss) | |
| I0525 02:28:00.647159 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00378059 (* 0.0909091 = 0.00034369 loss) | |
| I0525 02:28:00.647173 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00375345 (* 0.0909091 = 0.000341222 loss) | |
| I0525 02:28:00.647187 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00401738 (* 0.0909091 = 0.000365217 loss) | |
| I0525 02:28:00.647202 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00264488 (* 0.0909091 = 0.000240444 loss) | |
| I0525 02:28:00.647214 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00167358 (* 0.0909091 = 0.000152143 loss) | |
| I0525 02:28:00.647228 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:28:00.647238 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:28:00.647249 5272 solver.cpp:245] Train net output #149: total_confidence = 7.321e-07 | |
| I0525 02:28:00.647271 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.81173e-05 | |
| I0525 02:28:00.647286 5272 sgd_solver.cpp:106] Iteration 10500, lr = 0.001 | |
| I0525 02:28:27.982177 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.6898 > 30) by scale factor 0.946676 | |
| I0525 02:29:21.902336 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.3312 > 30) by scale factor 0.900059 | |
| I0525 02:30:11.175292 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.8269 > 30) by scale factor 0.861403 | |
| I0525 02:30:38.887442 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.3263 > 30) by scale factor 0.849225 | |
| I0525 02:30:55.048616 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.9548 > 30) by scale factor 0.969156 | |
| I0525 02:30:57.361604 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 121.946 > 30) by scale factor 0.24601 | |
| I0525 02:32:04.335583 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 64.8467 > 30) by scale factor 0.462629 | |
| I0525 02:33:31.359506 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.6863 > 30) by scale factor 0.864895 | |
| I0525 02:34:25.661705 5272 solver.cpp:229] Iteration 11000, loss = 10.4521 | |
| I0525 02:34:25.661849 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.06 | |
| I0525 02:34:25.661870 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 02:34:25.661885 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 02:34:25.661898 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 02:34:25.661911 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 02:34:25.661926 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 02:34:25.661938 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 02:34:25.661950 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 02:34:25.661963 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 02:34:25.661977 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 02:34:25.661989 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 02:34:25.662001 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 02:34:25.662014 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 02:34:25.662027 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 02:34:25.662039 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 02:34:25.662051 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:34:25.662063 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:34:25.662075 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:34:25.662087 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:34:25.662099 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:34:25.662111 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:34:25.662123 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:34:25.662134 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:34:25.662147 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.670455 | |
| I0525 02:34:25.662158 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.22 | |
| I0525 02:34:25.662174 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.30386 (* 0.3 = 0.991158 loss) | |
| I0525 02:34:25.662189 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.25399 (* 0.3 = 0.376196 loss) | |
| I0525 02:34:25.662204 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.27002 (* 0.0272727 = 0.0891824 loss) | |
| I0525 02:34:25.662217 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.33604 (* 0.0272727 = 0.090983 loss) | |
| I0525 02:34:25.662231 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.67329 (* 0.0272727 = 0.100181 loss) | |
| I0525 02:34:25.662245 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.06781 (* 0.0272727 = 0.0836675 loss) | |
| I0525 02:34:25.662259 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.86534 (* 0.0272727 = 0.0781455 loss) | |
| I0525 02:34:25.662273 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.12376 (* 0.0272727 = 0.0851934 loss) | |
| I0525 02:34:25.662287 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 0.839753 (* 0.0272727 = 0.0229024 loss) | |
| I0525 02:34:25.662302 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.433767 (* 0.0272727 = 0.01183 loss) | |
| I0525 02:34:25.662315 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.462224 (* 0.0272727 = 0.0126061 loss) | |
| I0525 02:34:25.662330 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.610093 (* 0.0272727 = 0.0166389 loss) | |
| I0525 02:34:25.662344 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.679484 (* 0.0272727 = 0.0185314 loss) | |
| I0525 02:34:25.662358 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.589567 (* 0.0272727 = 0.0160791 loss) | |
| I0525 02:34:25.662371 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.469102 (* 0.0272727 = 0.0127937 loss) | |
| I0525 02:34:25.662407 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.4895 (* 0.0272727 = 0.01335 loss) | |
| I0525 02:34:25.662422 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.215133 (* 0.0272727 = 0.00586727 loss) | |
| I0525 02:34:25.662437 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0876558 (* 0.0272727 = 0.00239061 loss) | |
| I0525 02:34:25.662451 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0503884 (* 0.0272727 = 0.00137423 loss) | |
| I0525 02:34:25.662466 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0269345 (* 0.0272727 = 0.000734577 loss) | |
| I0525 02:34:25.662480 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00856045 (* 0.0272727 = 0.000233467 loss) | |
| I0525 02:34:25.662494 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0116226 (* 0.0272727 = 0.000316981 loss) | |
| I0525 02:34:25.662508 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0152611 (* 0.0272727 = 0.000416211 loss) | |
| I0525 02:34:25.662523 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0148711 (* 0.0272727 = 0.000405576 loss) | |
| I0525 02:34:25.662535 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.06 | |
| I0525 02:34:25.662547 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 02:34:25.662560 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 02:34:25.662571 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 02:34:25.662583 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 02:34:25.662595 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 02:34:25.662607 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 02:34:25.662619 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 02:34:25.662631 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 02:34:25.662643 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 02:34:25.662655 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 02:34:25.662667 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 02:34:25.662678 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 02:34:25.662690 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 02:34:25.662703 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 02:34:25.662714 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:34:25.662726 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:34:25.662739 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:34:25.662750 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:34:25.662761 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:34:25.662773 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:34:25.662784 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:34:25.662796 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:34:25.662807 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.732955 | |
| I0525 02:34:25.662819 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.18 | |
| I0525 02:34:25.662833 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.42695 (* 0.3 = 1.02808 loss) | |
| I0525 02:34:25.662847 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.09081 (* 0.3 = 0.327243 loss) | |
| I0525 02:34:25.662864 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.26311 (* 0.0272727 = 0.088994 loss) | |
| I0525 02:34:25.662879 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.57417 (* 0.0272727 = 0.0974773 loss) | |
| I0525 02:34:25.662904 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.56997 (* 0.0272727 = 0.0973628 loss) | |
| I0525 02:34:25.662919 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.44516 (* 0.0272727 = 0.093959 loss) | |
| I0525 02:34:25.662936 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.99968 (* 0.0272727 = 0.0818096 loss) | |
| I0525 02:34:25.662950 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.04439 (* 0.0272727 = 0.0830289 loss) | |
| I0525 02:34:25.662963 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.22552 (* 0.0272727 = 0.0334234 loss) | |
| I0525 02:34:25.662977 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.378498 (* 0.0272727 = 0.0103227 loss) | |
| I0525 02:34:25.662992 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.697961 (* 0.0272727 = 0.0190353 loss) | |
| I0525 02:34:25.663005 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.497688 (* 0.0272727 = 0.0135733 loss) | |
| I0525 02:34:25.663019 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.522105 (* 0.0272727 = 0.0142392 loss) | |
| I0525 02:34:25.663033 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.571712 (* 0.0272727 = 0.0155921 loss) | |
| I0525 02:34:25.663048 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.426438 (* 0.0272727 = 0.0116301 loss) | |
| I0525 02:34:25.663061 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.539218 (* 0.0272727 = 0.014706 loss) | |
| I0525 02:34:25.663075 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.122262 (* 0.0272727 = 0.00333443 loss) | |
| I0525 02:34:25.663090 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0811688 (* 0.0272727 = 0.0022137 loss) | |
| I0525 02:34:25.663105 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0681039 (* 0.0272727 = 0.00185738 loss) | |
| I0525 02:34:25.663118 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0211841 (* 0.0272727 = 0.000577747 loss) | |
| I0525 02:34:25.663132 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0206871 (* 0.0272727 = 0.000564194 loss) | |
| I0525 02:34:25.663146 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0233919 (* 0.0272727 = 0.000637961 loss) | |
| I0525 02:34:25.663161 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0166734 (* 0.0272727 = 0.000454728 loss) | |
| I0525 02:34:25.663174 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0120139 (* 0.0272727 = 0.000327651 loss) | |
| I0525 02:34:25.663187 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.08 | |
| I0525 02:34:25.663198 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 02:34:25.663210 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 02:34:25.663223 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 02:34:25.663234 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 02:34:25.663246 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 02:34:25.663259 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 02:34:25.663270 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 02:34:25.663282 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 02:34:25.663295 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 02:34:25.663306 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 02:34:25.663318 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 02:34:25.663331 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 02:34:25.663342 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 02:34:25.663353 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 02:34:25.663365 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:34:25.663378 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:34:25.663398 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:34:25.663411 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:34:25.663422 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:34:25.663434 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:34:25.663446 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:34:25.663458 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:34:25.663470 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.738636 | |
| I0525 02:34:25.663482 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.16 | |
| I0525 02:34:25.663496 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.27506 (* 1 = 3.27506 loss) | |
| I0525 02:34:25.663509 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.05463 (* 1 = 1.05463 loss) | |
| I0525 02:34:25.663524 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.02914 (* 0.0909091 = 0.275376 loss) | |
| I0525 02:34:25.663534 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.3254 (* 0.0909091 = 0.302309 loss) | |
| I0525 02:34:25.663550 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.24472 (* 0.0909091 = 0.294975 loss) | |
| I0525 02:34:25.663565 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.85569 (* 0.0909091 = 0.259609 loss) | |
| I0525 02:34:25.663578 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.93663 (* 0.0909091 = 0.266966 loss) | |
| I0525 02:34:25.663592 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.01443 (* 0.0909091 = 0.27404 loss) | |
| I0525 02:34:25.663606 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.686808 (* 0.0909091 = 0.0624371 loss) | |
| I0525 02:34:25.663619 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.24189 (* 0.0909091 = 0.02199 loss) | |
| I0525 02:34:25.663633 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.528777 (* 0.0909091 = 0.0480707 loss) | |
| I0525 02:34:25.663647 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.297723 (* 0.0909091 = 0.0270657 loss) | |
| I0525 02:34:25.663661 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.459846 (* 0.0909091 = 0.0418042 loss) | |
| I0525 02:34:25.663676 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.560964 (* 0.0909091 = 0.0509967 loss) | |
| I0525 02:34:25.663689 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.405331 (* 0.0909091 = 0.0368483 loss) | |
| I0525 02:34:25.663703 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.445693 (* 0.0909091 = 0.0405175 loss) | |
| I0525 02:34:25.663718 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.188324 (* 0.0909091 = 0.0171204 loss) | |
| I0525 02:34:25.663733 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.126951 (* 0.0909091 = 0.011541 loss) | |
| I0525 02:34:25.663746 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0188411 (* 0.0909091 = 0.00171282 loss) | |
| I0525 02:34:25.663760 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0147675 (* 0.0909091 = 0.0013425 loss) | |
| I0525 02:34:25.663775 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0099363 (* 0.0909091 = 0.0009033 loss) | |
| I0525 02:34:25.663789 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00753917 (* 0.0909091 = 0.000685379 loss) | |
| I0525 02:34:25.663803 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00505063 (* 0.0909091 = 0.000459148 loss) | |
| I0525 02:34:25.663818 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00374615 (* 0.0909091 = 0.000340559 loss) | |
| I0525 02:34:25.663830 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:34:25.663842 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:34:25.663853 5272 solver.cpp:245] Train net output #149: total_confidence = 4.82095e-05 | |
| I0525 02:34:25.663874 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000520887 | |
| I0525 02:34:25.663888 5272 sgd_solver.cpp:106] Iteration 11000, lr = 0.001 | |
| I0525 02:35:24.522387 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8059 > 30) by scale factor 0.973839 | |
| I0525 02:37:30.804821 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 65.8581 > 30) by scale factor 0.455525 | |
| I0525 02:40:18.634148 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.6774 > 30) by scale factor 0.977918 | |
| I0525 02:40:25.571470 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.0927 > 30) by scale factor 0.879954 | |
| I0525 02:40:50.595798 5272 solver.cpp:229] Iteration 11500, loss = 10.3291 | |
| I0525 02:40:50.595916 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0784314 | |
| I0525 02:40:50.595937 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 02:40:50.595950 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 02:40:50.595963 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 02:40:50.595975 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 02:40:50.595988 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 02:40:50.596000 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 02:40:50.596014 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 02:40:50.596026 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 02:40:50.596038 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 02:40:50.596051 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 02:40:50.596065 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 02:40:50.596081 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 02:40:50.596103 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 02:40:50.596122 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:40:50.596138 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:40:50.596151 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:40:50.596163 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:40:50.596174 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:40:50.596186 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:40:50.596199 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:40:50.596210 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:40:50.596221 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:40:50.596233 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.704545 | |
| I0525 02:40:50.596253 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.176471 | |
| I0525 02:40:50.596279 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.42657 (* 0.3 = 1.02797 loss) | |
| I0525 02:40:50.596295 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.26096 (* 0.3 = 0.378289 loss) | |
| I0525 02:40:50.596309 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.94878 (* 0.0272727 = 0.107694 loss) | |
| I0525 02:40:50.596323 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.57616 (* 0.0272727 = 0.0975317 loss) | |
| I0525 02:40:50.596338 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 4.17248 (* 0.0272727 = 0.113795 loss) | |
| I0525 02:40:50.596351 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.35388 (* 0.0272727 = 0.0914694 loss) | |
| I0525 02:40:50.596365 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.04976 (* 0.0272727 = 0.0831753 loss) | |
| I0525 02:40:50.596379 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.95395 (* 0.0272727 = 0.0805624 loss) | |
| I0525 02:40:50.596393 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.25185 (* 0.0272727 = 0.0614141 loss) | |
| I0525 02:40:50.596407 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.707424 (* 0.0272727 = 0.0192934 loss) | |
| I0525 02:40:50.596421 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.477193 (* 0.0272727 = 0.0130144 loss) | |
| I0525 02:40:50.596436 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.782575 (* 0.0272727 = 0.0213429 loss) | |
| I0525 02:40:50.596449 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.475722 (* 0.0272727 = 0.0129742 loss) | |
| I0525 02:40:50.596463 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.666057 (* 0.0272727 = 0.0181652 loss) | |
| I0525 02:40:50.596478 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0716908 (* 0.0272727 = 0.0019552 loss) | |
| I0525 02:40:50.596513 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0513454 (* 0.0272727 = 0.00140033 loss) | |
| I0525 02:40:50.596529 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0256102 (* 0.0272727 = 0.000698461 loss) | |
| I0525 02:40:50.596544 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.027763 (* 0.0272727 = 0.000757173 loss) | |
| I0525 02:40:50.596557 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0154844 (* 0.0272727 = 0.000422302 loss) | |
| I0525 02:40:50.596571 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0216072 (* 0.0272727 = 0.000589287 loss) | |
| I0525 02:40:50.596585 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0148684 (* 0.0272727 = 0.000405503 loss) | |
| I0525 02:40:50.596599 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0117548 (* 0.0272727 = 0.000320587 loss) | |
| I0525 02:40:50.596613 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0240704 (* 0.0272727 = 0.000656465 loss) | |
| I0525 02:40:50.596627 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00909316 (* 0.0272727 = 0.000247995 loss) | |
| I0525 02:40:50.596639 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0784314 | |
| I0525 02:40:50.596652 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 02:40:50.596664 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 02:40:50.596676 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 02:40:50.596688 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 02:40:50.596700 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 02:40:50.596712 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 02:40:50.596724 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 02:40:50.596736 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 02:40:50.596748 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 02:40:50.596760 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 02:40:50.596771 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 02:40:50.596783 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 02:40:50.596797 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 02:40:50.596810 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:40:50.596822 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:40:50.596834 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:40:50.596845 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:40:50.596858 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:40:50.596868 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:40:50.596880 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:40:50.596892 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:40:50.596904 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:40:50.596915 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.715909 | |
| I0525 02:40:50.596926 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.137255 | |
| I0525 02:40:50.596940 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.35836 (* 0.3 = 1.00751 loss) | |
| I0525 02:40:50.596954 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.19902 (* 0.3 = 0.359705 loss) | |
| I0525 02:40:50.596968 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 4.31821 (* 0.0272727 = 0.117769 loss) | |
| I0525 02:40:50.596982 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.82222 (* 0.0272727 = 0.104242 loss) | |
| I0525 02:40:50.597007 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 4.08618 (* 0.0272727 = 0.111441 loss) | |
| I0525 02:40:50.597021 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.14727 (* 0.0272727 = 0.0858346 loss) | |
| I0525 02:40:50.597036 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.03893 (* 0.0272727 = 0.0828798 loss) | |
| I0525 02:40:50.597061 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.19486 (* 0.0272727 = 0.0871327 loss) | |
| I0525 02:40:50.597087 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.01338 (* 0.0272727 = 0.0549104 loss) | |
| I0525 02:40:50.597105 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.553043 (* 0.0272727 = 0.015083 loss) | |
| I0525 02:40:50.597132 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.376881 (* 0.0272727 = 0.0102786 loss) | |
| I0525 02:40:50.597149 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.535015 (* 0.0272727 = 0.0145913 loss) | |
| I0525 02:40:50.597163 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.445996 (* 0.0272727 = 0.0121635 loss) | |
| I0525 02:40:50.597177 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.612891 (* 0.0272727 = 0.0167152 loss) | |
| I0525 02:40:50.597195 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0930959 (* 0.0272727 = 0.00253898 loss) | |
| I0525 02:40:50.597210 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0614963 (* 0.0272727 = 0.00167717 loss) | |
| I0525 02:40:50.597224 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0474584 (* 0.0272727 = 0.00129432 loss) | |
| I0525 02:40:50.597239 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0264458 (* 0.0272727 = 0.000721249 loss) | |
| I0525 02:40:50.597252 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0189328 (* 0.0272727 = 0.00051635 loss) | |
| I0525 02:40:50.597262 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00799538 (* 0.0272727 = 0.000218056 loss) | |
| I0525 02:40:50.597277 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00849798 (* 0.0272727 = 0.000231763 loss) | |
| I0525 02:40:50.597292 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0117907 (* 0.0272727 = 0.000321564 loss) | |
| I0525 02:40:50.597306 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00699201 (* 0.0272727 = 0.000190691 loss) | |
| I0525 02:40:50.597319 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00597752 (* 0.0272727 = 0.000163023 loss) | |
| I0525 02:40:50.597332 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0588235 | |
| I0525 02:40:50.597343 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 02:40:50.597355 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0525 02:40:50.597368 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 02:40:50.597379 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 02:40:50.597390 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 02:40:50.597403 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 02:40:50.597414 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 02:40:50.597425 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 02:40:50.597437 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 02:40:50.597450 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 02:40:50.597460 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 02:40:50.597472 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 02:40:50.597483 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 02:40:50.597496 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:40:50.597506 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:40:50.597518 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:40:50.597543 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:40:50.597554 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:40:50.597566 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:40:50.597578 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:40:50.597589 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:40:50.597600 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:40:50.597612 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.715909 | |
| I0525 02:40:50.597623 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.176471 | |
| I0525 02:40:50.597637 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.33208 (* 1 = 3.33208 loss) | |
| I0525 02:40:50.597651 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.08258 (* 1 = 1.08258 loss) | |
| I0525 02:40:50.597666 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 4.04539 (* 0.0909091 = 0.367763 loss) | |
| I0525 02:40:50.597678 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.44273 (* 0.0909091 = 0.312976 loss) | |
| I0525 02:40:50.597692 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.51712 (* 0.0909091 = 0.319738 loss) | |
| I0525 02:40:50.597705 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.66002 (* 0.0909091 = 0.24182 loss) | |
| I0525 02:40:50.597719 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.60911 (* 0.0909091 = 0.237192 loss) | |
| I0525 02:40:50.597733 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.89707 (* 0.0909091 = 0.26337 loss) | |
| I0525 02:40:50.597746 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.94152 (* 0.0909091 = 0.176502 loss) | |
| I0525 02:40:50.597760 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.544194 (* 0.0909091 = 0.0494722 loss) | |
| I0525 02:40:50.597774 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.295606 (* 0.0909091 = 0.0268733 loss) | |
| I0525 02:40:50.597787 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.669447 (* 0.0909091 = 0.0608588 loss) | |
| I0525 02:40:50.597801 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.493635 (* 0.0909091 = 0.0448759 loss) | |
| I0525 02:40:50.597815 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.678522 (* 0.0909091 = 0.0616839 loss) | |
| I0525 02:40:50.597828 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0271028 (* 0.0909091 = 0.00246389 loss) | |
| I0525 02:40:50.597843 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0324675 (* 0.0909091 = 0.00295159 loss) | |
| I0525 02:40:50.597861 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0100096 (* 0.0909091 = 0.000909964 loss) | |
| I0525 02:40:50.597874 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00932357 (* 0.0909091 = 0.000847598 loss) | |
| I0525 02:40:50.597889 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00817514 (* 0.0909091 = 0.000743194 loss) | |
| I0525 02:40:50.597903 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00484145 (* 0.0909091 = 0.000440132 loss) | |
| I0525 02:40:50.597916 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00303621 (* 0.0909091 = 0.000276019 loss) | |
| I0525 02:40:50.597930 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00211655 (* 0.0909091 = 0.000192413 loss) | |
| I0525 02:40:50.597944 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00201174 (* 0.0909091 = 0.000182886 loss) | |
| I0525 02:40:50.597959 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00108973 (* 0.0909091 = 9.90664e-05 loss) | |
| I0525 02:40:50.597967 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:40:50.597980 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:40:50.597991 5272 solver.cpp:245] Train net output #149: total_confidence = 1.21884e-06 | |
| I0525 02:40:50.598013 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.53115e-05 | |
| I0525 02:40:50.598028 5272 sgd_solver.cpp:106] Iteration 11500, lr = 0.001 | |
| I0525 02:42:07.130846 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.4887 > 30) by scale factor 0.869849 | |
| I0525 02:47:15.463932 5272 solver.cpp:229] Iteration 12000, loss = 10.3201 | |
| I0525 02:47:15.464093 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0508475 | |
| I0525 02:47:15.464115 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 02:47:15.464134 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 02:47:15.464154 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 02:47:15.464167 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 02:47:15.464182 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 02:47:15.464201 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 02:47:15.464215 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 02:47:15.464228 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.5 | |
| I0525 02:47:15.464241 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.625 | |
| I0525 02:47:15.464253 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 02:47:15.464267 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 02:47:15.464278 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 02:47:15.464290 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 02:47:15.464303 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 02:47:15.464314 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 02:47:15.464326 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 02:47:15.464339 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 02:47:15.464349 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 02:47:15.464361 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:47:15.464373 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:47:15.464385 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:47:15.464396 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:47:15.464408 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.681818 | |
| I0525 02:47:15.464421 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.186441 | |
| I0525 02:47:15.464437 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.26478 (* 0.3 = 0.979435 loss) | |
| I0525 02:47:15.464450 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.2857 (* 0.3 = 0.38571 loss) | |
| I0525 02:47:15.464465 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.42404 (* 0.0272727 = 0.093383 loss) | |
| I0525 02:47:15.464479 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.70486 (* 0.0272727 = 0.101042 loss) | |
| I0525 02:47:15.464495 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.1733 (* 0.0272727 = 0.0865445 loss) | |
| I0525 02:47:15.464509 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.26134 (* 0.0272727 = 0.0889457 loss) | |
| I0525 02:47:15.464524 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.68376 (* 0.0272727 = 0.0731934 loss) | |
| I0525 02:47:15.464537 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.64407 (* 0.0272727 = 0.072111 loss) | |
| I0525 02:47:15.464550 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.20124 (* 0.0272727 = 0.0600339 loss) | |
| I0525 02:47:15.464565 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.983 (* 0.0272727 = 0.0540817 loss) | |
| I0525 02:47:15.464578 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.56478 (* 0.0272727 = 0.0426759 loss) | |
| I0525 02:47:15.464592 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.781675 (* 0.0272727 = 0.0213184 loss) | |
| I0525 02:47:15.464607 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.646583 (* 0.0272727 = 0.0176341 loss) | |
| I0525 02:47:15.464620 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.63094 (* 0.0272727 = 0.0172075 loss) | |
| I0525 02:47:15.464634 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.825503 (* 0.0272727 = 0.0225137 loss) | |
| I0525 02:47:15.464670 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0709609 (* 0.0272727 = 0.0019353 loss) | |
| I0525 02:47:15.464686 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0666526 (* 0.0272727 = 0.0018178 loss) | |
| I0525 02:47:15.464700 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0407651 (* 0.0272727 = 0.00111178 loss) | |
| I0525 02:47:15.464715 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0315376 (* 0.0272727 = 0.000860115 loss) | |
| I0525 02:47:15.464730 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0160632 (* 0.0272727 = 0.000438086 loss) | |
| I0525 02:47:15.464743 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.013271 (* 0.0272727 = 0.000361937 loss) | |
| I0525 02:47:15.464757 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0114999 (* 0.0272727 = 0.000313633 loss) | |
| I0525 02:47:15.464771 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0105122 (* 0.0272727 = 0.000286697 loss) | |
| I0525 02:47:15.464786 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00871918 (* 0.0272727 = 0.000237796 loss) | |
| I0525 02:47:15.464797 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0847458 | |
| I0525 02:47:15.464809 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.375 | |
| I0525 02:47:15.464823 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 02:47:15.464834 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 02:47:15.464846 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 02:47:15.464857 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 02:47:15.464870 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 02:47:15.464885 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 02:47:15.464897 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.5 | |
| I0525 02:47:15.464910 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.625 | |
| I0525 02:47:15.464925 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 02:47:15.464946 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 02:47:15.464959 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 02:47:15.464972 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 02:47:15.464988 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 02:47:15.465004 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 02:47:15.465018 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 02:47:15.465029 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 02:47:15.465040 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 02:47:15.465052 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:47:15.465064 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:47:15.465075 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:47:15.465086 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:47:15.465098 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0525 02:47:15.465127 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.322034 | |
| I0525 02:47:15.465142 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.13707 (* 0.3 = 0.941122 loss) | |
| I0525 02:47:15.465157 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.22715 (* 0.3 = 0.368146 loss) | |
| I0525 02:47:15.465169 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.79942 (* 0.0272727 = 0.0763478 loss) | |
| I0525 02:47:15.465183 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.62237 (* 0.0272727 = 0.0987919 loss) | |
| I0525 02:47:15.465211 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.0321 (* 0.0272727 = 0.0826935 loss) | |
| I0525 02:47:15.465226 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.50749 (* 0.0272727 = 0.0956587 loss) | |
| I0525 02:47:15.465240 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.82561 (* 0.0272727 = 0.077062 loss) | |
| I0525 02:47:15.465255 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.72742 (* 0.0272727 = 0.0743841 loss) | |
| I0525 02:47:15.465267 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.42663 (* 0.0272727 = 0.0661809 loss) | |
| I0525 02:47:15.465282 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 2.20712 (* 0.0272727 = 0.0601941 loss) | |
| I0525 02:47:15.465296 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.21751 (* 0.0272727 = 0.0332047 loss) | |
| I0525 02:47:15.465309 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.652865 (* 0.0272727 = 0.0178054 loss) | |
| I0525 02:47:15.465323 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.744584 (* 0.0272727 = 0.0203068 loss) | |
| I0525 02:47:15.465337 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.647154 (* 0.0272727 = 0.0176496 loss) | |
| I0525 02:47:15.465351 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.984918 (* 0.0272727 = 0.0268614 loss) | |
| I0525 02:47:15.465365 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0356069 (* 0.0272727 = 0.000971097 loss) | |
| I0525 02:47:15.465379 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0473983 (* 0.0272727 = 0.00129268 loss) | |
| I0525 02:47:15.465394 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.010702 (* 0.0272727 = 0.000291873 loss) | |
| I0525 02:47:15.465409 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00669342 (* 0.0272727 = 0.000182548 loss) | |
| I0525 02:47:15.465422 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00251353 (* 0.0272727 = 6.85508e-05 loss) | |
| I0525 02:47:15.465436 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00257808 (* 0.0272727 = 7.03114e-05 loss) | |
| I0525 02:47:15.465451 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00218635 (* 0.0272727 = 5.96278e-05 loss) | |
| I0525 02:47:15.465464 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00195691 (* 0.0272727 = 5.33702e-05 loss) | |
| I0525 02:47:15.465477 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00258035 (* 0.0272727 = 7.03732e-05 loss) | |
| I0525 02:47:15.465490 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.152542 | |
| I0525 02:47:15.465502 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 02:47:15.465514 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 02:47:15.465528 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0525 02:47:15.465539 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 02:47:15.465550 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 02:47:15.465562 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 02:47:15.465574 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 02:47:15.465586 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.5 | |
| I0525 02:47:15.465597 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0525 02:47:15.465610 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 02:47:15.465621 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 02:47:15.465633 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 02:47:15.465646 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 02:47:15.465657 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 02:47:15.465668 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 02:47:15.465680 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 02:47:15.465698 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 02:47:15.465716 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 02:47:15.465734 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:47:15.465747 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:47:15.465764 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:47:15.465778 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:47:15.465790 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0525 02:47:15.465802 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.338983 | |
| I0525 02:47:15.465817 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.92974 (* 1 = 2.92974 loss) | |
| I0525 02:47:15.465831 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.16516 (* 1 = 1.16516 loss) | |
| I0525 02:47:15.465844 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.46726 (* 0.0909091 = 0.224297 loss) | |
| I0525 02:47:15.465858 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.39848 (* 0.0909091 = 0.308953 loss) | |
| I0525 02:47:15.465873 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.69394 (* 0.0909091 = 0.244904 loss) | |
| I0525 02:47:15.465886 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.86157 (* 0.0909091 = 0.260143 loss) | |
| I0525 02:47:15.465899 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.5146 (* 0.0909091 = 0.2286 loss) | |
| I0525 02:47:15.465914 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.9573 (* 0.0909091 = 0.177936 loss) | |
| I0525 02:47:15.465930 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.77448 (* 0.0909091 = 0.161316 loss) | |
| I0525 02:47:15.465945 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.72743 (* 0.0909091 = 0.157039 loss) | |
| I0525 02:47:15.465958 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.984024 (* 0.0909091 = 0.0894567 loss) | |
| I0525 02:47:15.465972 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.599245 (* 0.0909091 = 0.0544768 loss) | |
| I0525 02:47:15.465986 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.526548 (* 0.0909091 = 0.047868 loss) | |
| I0525 02:47:15.466001 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.405981 (* 0.0909091 = 0.0369074 loss) | |
| I0525 02:47:15.466013 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.597492 (* 0.0909091 = 0.0543174 loss) | |
| I0525 02:47:15.466028 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.108092 (* 0.0909091 = 0.00982653 loss) | |
| I0525 02:47:15.466042 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0813336 (* 0.0909091 = 0.00739396 loss) | |
| I0525 02:47:15.466056 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0197809 (* 0.0909091 = 0.00179826 loss) | |
| I0525 02:47:15.466070 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0109689 (* 0.0909091 = 0.000997171 loss) | |
| I0525 02:47:15.466084 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00615124 (* 0.0909091 = 0.000559204 loss) | |
| I0525 02:47:15.466097 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00338673 (* 0.0909091 = 0.000307884 loss) | |
| I0525 02:47:15.466111 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00114281 (* 0.0909091 = 0.000103892 loss) | |
| I0525 02:47:15.466125 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000616909 (* 0.0909091 = 5.60826e-05 loss) | |
| I0525 02:47:15.466138 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000255305 (* 0.0909091 = 2.32095e-05 loss) | |
| I0525 02:47:15.466150 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:47:15.466166 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:47:15.466186 5272 solver.cpp:245] Train net output #149: total_confidence = 1.91328e-05 | |
| I0525 02:47:15.466212 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00019134 | |
| I0525 02:47:15.466233 5272 sgd_solver.cpp:106] Iteration 12000, lr = 0.001 | |
| I0525 02:47:38.932745 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 71.5104 > 30) by scale factor 0.419519 | |
| I0525 02:49:08.235553 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.1562 > 30) by scale factor 0.853334 | |
| I0525 02:50:22.878186 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 61.7182 > 30) by scale factor 0.48608 | |
| I0525 02:50:45.199411 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 51.4936 > 30) by scale factor 0.582597 | |
| I0525 02:51:41.361454 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.1841 > 30) by scale factor 0.829092 | |
| I0525 02:53:36.006780 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.3767 > 30) by scale factor 0.80264 | |
| I0525 02:53:40.262481 5272 solver.cpp:229] Iteration 12500, loss = 10.2808 | |
| I0525 02:53:40.262545 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.137931 | |
| I0525 02:53:40.262563 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 02:53:40.262583 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0525 02:53:40.262609 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 02:53:40.262625 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 02:53:40.262639 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 02:53:40.262651 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 02:53:40.262663 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 02:53:40.262676 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 02:53:40.262689 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 02:53:40.262701 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 02:53:40.262715 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 02:53:40.262727 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 02:53:40.262747 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 02:53:40.262773 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 02:53:40.262787 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 02:53:40.262800 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 02:53:40.262812 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0525 02:53:40.262825 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0525 02:53:40.262836 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 02:53:40.262850 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 02:53:40.262861 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 02:53:40.262872 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 02:53:40.262884 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.704545 | |
| I0525 02:53:40.262897 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.362069 | |
| I0525 02:53:40.262912 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.06629 (* 0.3 = 0.919887 loss) | |
| I0525 02:53:40.262928 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.15324 (* 0.3 = 0.345972 loss) | |
| I0525 02:53:40.262943 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.76091 (* 0.0272727 = 0.0752975 loss) | |
| I0525 02:53:40.262956 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.20251 (* 0.0272727 = 0.0873412 loss) | |
| I0525 02:53:40.262970 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.00768 (* 0.0272727 = 0.0820275 loss) | |
| I0525 02:53:40.262984 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.12238 (* 0.0272727 = 0.0851557 loss) | |
| I0525 02:53:40.262998 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.87431 (* 0.0272727 = 0.0783902 loss) | |
| I0525 02:53:40.263012 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.57214 (* 0.0272727 = 0.0701492 loss) | |
| I0525 02:53:40.263025 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.92674 (* 0.0272727 = 0.0798202 loss) | |
| I0525 02:53:40.263041 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.2146 (* 0.0272727 = 0.00585272 loss) | |
| I0525 02:53:40.263054 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.561619 (* 0.0272727 = 0.0153169 loss) | |
| I0525 02:53:40.263068 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.467 (* 0.0272727 = 0.0127364 loss) | |
| I0525 02:53:40.263082 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.722842 (* 0.0272727 = 0.0197139 loss) | |
| I0525 02:53:40.263098 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.457214 (* 0.0272727 = 0.0124695 loss) | |
| I0525 02:53:40.263144 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.501079 (* 0.0272727 = 0.0136658 loss) | |
| I0525 02:53:40.263159 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.625255 (* 0.0272727 = 0.0170524 loss) | |
| I0525 02:53:40.263175 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.657307 (* 0.0272727 = 0.0179266 loss) | |
| I0525 02:53:40.263188 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 1.02872 (* 0.0272727 = 0.0280559 loss) | |
| I0525 02:53:40.263202 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 1.0249 (* 0.0272727 = 0.0279517 loss) | |
| I0525 02:53:40.263216 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.919519 (* 0.0272727 = 0.0250778 loss) | |
| I0525 02:53:40.263231 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00190191 (* 0.0272727 = 5.18703e-05 loss) | |
| I0525 02:53:40.263245 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00328753 (* 0.0272727 = 8.96599e-05 loss) | |
| I0525 02:53:40.263259 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00113713 (* 0.0272727 = 3.10127e-05 loss) | |
| I0525 02:53:40.263273 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000949502 (* 0.0272727 = 2.58955e-05 loss) | |
| I0525 02:53:40.263286 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0689655 | |
| I0525 02:53:40.263298 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 02:53:40.263311 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.375 | |
| I0525 02:53:40.263322 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 02:53:40.263334 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 02:53:40.263347 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 02:53:40.263360 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 02:53:40.263371 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 02:53:40.263382 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 02:53:40.263394 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 02:53:40.263406 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 02:53:40.263418 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 02:53:40.263430 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 02:53:40.263442 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 02:53:40.263453 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 02:53:40.263465 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 02:53:40.263478 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 02:53:40.263489 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0525 02:53:40.263501 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0525 02:53:40.263512 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 02:53:40.263525 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 02:53:40.263536 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 02:53:40.263547 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 02:53:40.263559 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0525 02:53:40.263571 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.258621 | |
| I0525 02:53:40.263586 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.07766 (* 0.3 = 0.923298 loss) | |
| I0525 02:53:40.263599 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.12056 (* 0.3 = 0.336167 loss) | |
| I0525 02:53:40.263613 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.3003 (* 0.0272727 = 0.0900082 loss) | |
| I0525 02:53:40.263638 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.08696 (* 0.0272727 = 0.0841899 loss) | |
| I0525 02:53:40.263653 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.70705 (* 0.0272727 = 0.101101 loss) | |
| I0525 02:53:40.263667 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 2.8739 (* 0.0272727 = 0.0783792 loss) | |
| I0525 02:53:40.263681 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.63683 (* 0.0272727 = 0.0719135 loss) | |
| I0525 02:53:40.263695 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.97925 (* 0.0272727 = 0.0812522 loss) | |
| I0525 02:53:40.263710 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.14341 (* 0.0272727 = 0.0584567 loss) | |
| I0525 02:53:40.263723 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.441174 (* 0.0272727 = 0.012032 loss) | |
| I0525 02:53:40.263737 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.825685 (* 0.0272727 = 0.0225187 loss) | |
| I0525 02:53:40.263751 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.503092 (* 0.0272727 = 0.0137207 loss) | |
| I0525 02:53:40.263767 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.782765 (* 0.0272727 = 0.0213481 loss) | |
| I0525 02:53:40.263793 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.479955 (* 0.0272727 = 0.0130897 loss) | |
| I0525 02:53:40.263820 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.834026 (* 0.0272727 = 0.0227462 loss) | |
| I0525 02:53:40.263835 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.660875 (* 0.0272727 = 0.0180239 loss) | |
| I0525 02:53:40.263851 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.595649 (* 0.0272727 = 0.016245 loss) | |
| I0525 02:53:40.263861 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.905326 (* 0.0272727 = 0.0246907 loss) | |
| I0525 02:53:40.263876 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.886569 (* 0.0272727 = 0.0241792 loss) | |
| I0525 02:53:40.263890 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 1.05707 (* 0.0272727 = 0.0288291 loss) | |
| I0525 02:53:40.263905 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0162294 (* 0.0272727 = 0.000442619 loss) | |
| I0525 02:53:40.263919 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.015836 (* 0.0272727 = 0.000431891 loss) | |
| I0525 02:53:40.263933 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0127336 (* 0.0272727 = 0.000347281 loss) | |
| I0525 02:53:40.263947 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00798731 (* 0.0272727 = 0.000217836 loss) | |
| I0525 02:53:40.263959 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.137931 | |
| I0525 02:53:40.263972 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 02:53:40.263984 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 02:53:40.263995 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 02:53:40.264008 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 02:53:40.264020 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 02:53:40.264032 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 02:53:40.264044 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 02:53:40.264056 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 02:53:40.264067 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 02:53:40.264080 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 02:53:40.264091 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 02:53:40.264103 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 02:53:40.264116 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 02:53:40.264127 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 02:53:40.264138 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 02:53:40.264163 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 02:53:40.264175 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0525 02:53:40.264188 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0525 02:53:40.264199 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 02:53:40.264211 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 02:53:40.264224 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 02:53:40.264235 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 02:53:40.264246 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.704545 | |
| I0525 02:53:40.264258 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.310345 | |
| I0525 02:53:40.264272 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.92147 (* 1 = 2.92147 loss) | |
| I0525 02:53:40.264286 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.10729 (* 1 = 1.10729 loss) | |
| I0525 02:53:40.264302 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.66969 (* 0.0909091 = 0.242699 loss) | |
| I0525 02:53:40.264315 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.92728 (* 0.0909091 = 0.266116 loss) | |
| I0525 02:53:40.264329 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.32159 (* 0.0909091 = 0.301962 loss) | |
| I0525 02:53:40.264343 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.80263 (* 0.0909091 = 0.254784 loss) | |
| I0525 02:53:40.264356 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.48772 (* 0.0909091 = 0.226156 loss) | |
| I0525 02:53:40.264370 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.42969 (* 0.0909091 = 0.220881 loss) | |
| I0525 02:53:40.264384 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.32776 (* 0.0909091 = 0.211614 loss) | |
| I0525 02:53:40.264397 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.277731 (* 0.0909091 = 0.0252483 loss) | |
| I0525 02:53:40.264411 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.589188 (* 0.0909091 = 0.0535625 loss) | |
| I0525 02:53:40.264425 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.374044 (* 0.0909091 = 0.034004 loss) | |
| I0525 02:53:40.264439 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.423378 (* 0.0909091 = 0.0384889 loss) | |
| I0525 02:53:40.264453 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.476018 (* 0.0909091 = 0.0432744 loss) | |
| I0525 02:53:40.264467 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.37512 (* 0.0909091 = 0.0341018 loss) | |
| I0525 02:53:40.264480 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.573537 (* 0.0909091 = 0.0521397 loss) | |
| I0525 02:53:40.264494 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.31151 (* 0.0909091 = 0.0283191 loss) | |
| I0525 02:53:40.264508 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.629892 (* 0.0909091 = 0.0572629 loss) | |
| I0525 02:53:40.264521 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.647971 (* 0.0909091 = 0.0589065 loss) | |
| I0525 02:53:40.264535 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.715721 (* 0.0909091 = 0.0650655 loss) | |
| I0525 02:53:40.264549 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00276362 (* 0.0909091 = 0.000251239 loss) | |
| I0525 02:53:40.264564 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00359171 (* 0.0909091 = 0.000326519 loss) | |
| I0525 02:53:40.264577 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000527144 (* 0.0909091 = 4.79222e-05 loss) | |
| I0525 02:53:40.264591 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000319462 (* 0.0909091 = 2.9042e-05 loss) | |
| I0525 02:53:40.264603 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 02:53:40.264616 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 02:53:40.264632 5272 solver.cpp:245] Train net output #149: total_confidence = 2.72203e-06 | |
| I0525 02:53:40.264647 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000281044 | |
| I0525 02:53:40.264660 5272 sgd_solver.cpp:106] Iteration 12500, lr = 0.001 | |
| I0525 02:56:09.888660 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.7626 > 30) by scale factor 0.888558 | |
| I0525 03:00:04.984876 5272 solver.cpp:229] Iteration 13000, loss = 10.2016 | |
| I0525 03:00:04.985003 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0434783 | |
| I0525 03:00:04.985023 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 03:00:04.985038 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 03:00:04.985049 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 03:00:04.985062 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 03:00:04.985075 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 03:00:04.985088 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 03:00:04.985101 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 03:00:04.985115 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 03:00:04.985139 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:00:04.985153 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:00:04.985165 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:00:04.985178 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:00:04.985190 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:00:04.985203 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:00:04.985213 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:00:04.985226 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:00:04.985239 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:00:04.985249 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:00:04.985261 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:00:04.985273 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:00:04.985285 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:00:04.985297 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:00:04.985309 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.75 | |
| I0525 03:00:04.985322 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.217391 | |
| I0525 03:00:04.985339 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.13467 (* 0.3 = 0.940402 loss) | |
| I0525 03:00:04.985353 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.874241 (* 0.3 = 0.262272 loss) | |
| I0525 03:00:04.985368 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.46684 (* 0.0272727 = 0.0945503 loss) | |
| I0525 03:00:04.985383 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.22572 (* 0.0272727 = 0.0879742 loss) | |
| I0525 03:00:04.985396 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.37468 (* 0.0272727 = 0.0920368 loss) | |
| I0525 03:00:04.985410 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 4.00705 (* 0.0272727 = 0.109283 loss) | |
| I0525 03:00:04.985424 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.38 (* 0.0272727 = 0.064909 loss) | |
| I0525 03:00:04.985438 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.70713 (* 0.0272727 = 0.0738309 loss) | |
| I0525 03:00:04.985452 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.99212 (* 0.0272727 = 0.0543305 loss) | |
| I0525 03:00:04.985466 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.966078 (* 0.0272727 = 0.0263476 loss) | |
| I0525 03:00:04.985481 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.00874336 (* 0.0272727 = 0.000238455 loss) | |
| I0525 03:00:04.985494 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00450344 (* 0.0272727 = 0.000122821 loss) | |
| I0525 03:00:04.985509 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.00569316 (* 0.0272727 = 0.000155268 loss) | |
| I0525 03:00:04.985523 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00503099 (* 0.0272727 = 0.000137209 loss) | |
| I0525 03:00:04.985538 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00505718 (* 0.0272727 = 0.000137923 loss) | |
| I0525 03:00:04.985570 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00486754 (* 0.0272727 = 0.000132751 loss) | |
| I0525 03:00:04.985586 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00363545 (* 0.0272727 = 9.91486e-05 loss) | |
| I0525 03:00:04.985600 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00519197 (* 0.0272727 = 0.000141599 loss) | |
| I0525 03:00:04.985615 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00338695 (* 0.0272727 = 9.23715e-05 loss) | |
| I0525 03:00:04.985630 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0021278 (* 0.0272727 = 5.80308e-05 loss) | |
| I0525 03:00:04.985643 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00233622 (* 0.0272727 = 6.37151e-05 loss) | |
| I0525 03:00:04.985657 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00277386 (* 0.0272727 = 7.56508e-05 loss) | |
| I0525 03:00:04.985671 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0016814 (* 0.0272727 = 4.58564e-05 loss) | |
| I0525 03:00:04.985685 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00283452 (* 0.0272727 = 7.7305e-05 loss) | |
| I0525 03:00:04.985698 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0869565 | |
| I0525 03:00:04.985710 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 03:00:04.985723 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 03:00:04.985734 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 03:00:04.985745 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 03:00:04.985757 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 03:00:04.985769 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 03:00:04.985780 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 03:00:04.985792 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 03:00:04.985805 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:00:04.985816 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:00:04.985827 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:00:04.985838 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:00:04.985849 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:00:04.985860 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:00:04.985872 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:00:04.985888 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:00:04.985899 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:00:04.985910 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:00:04.985923 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:00:04.985934 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:00:04.985945 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:00:04.985956 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:00:04.985967 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0525 03:00:04.985980 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.217391 | |
| I0525 03:00:04.985993 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.01777 (* 0.3 = 0.905332 loss) | |
| I0525 03:00:04.986007 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.875113 (* 0.3 = 0.262534 loss) | |
| I0525 03:00:04.986024 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.21462 (* 0.0272727 = 0.0876714 loss) | |
| I0525 03:00:04.986039 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.96735 (* 0.0272727 = 0.0809277 loss) | |
| I0525 03:00:04.986063 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.54991 (* 0.0272727 = 0.0968156 loss) | |
| I0525 03:00:04.986078 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.46287 (* 0.0272727 = 0.0944419 loss) | |
| I0525 03:00:04.986093 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.88675 (* 0.0272727 = 0.0787295 loss) | |
| I0525 03:00:04.986105 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.34243 (* 0.0272727 = 0.0638846 loss) | |
| I0525 03:00:04.986119 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.28131 (* 0.0272727 = 0.0622175 loss) | |
| I0525 03:00:04.986132 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.0257 (* 0.0272727 = 0.0279736 loss) | |
| I0525 03:00:04.986147 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0211147 (* 0.0272727 = 0.000575854 loss) | |
| I0525 03:00:04.986160 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0116302 (* 0.0272727 = 0.000317188 loss) | |
| I0525 03:00:04.986174 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0094277 (* 0.0272727 = 0.000257119 loss) | |
| I0525 03:00:04.986188 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0111263 (* 0.0272727 = 0.000303445 loss) | |
| I0525 03:00:04.986202 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00945274 (* 0.0272727 = 0.000257802 loss) | |
| I0525 03:00:04.986217 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00796373 (* 0.0272727 = 0.000217193 loss) | |
| I0525 03:00:04.986230 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00867668 (* 0.0272727 = 0.000236637 loss) | |
| I0525 03:00:04.986245 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00412891 (* 0.0272727 = 0.000112607 loss) | |
| I0525 03:00:04.986259 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0044009 (* 0.0272727 = 0.000120025 loss) | |
| I0525 03:00:04.986274 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00244736 (* 0.0272727 = 6.67463e-05 loss) | |
| I0525 03:00:04.986286 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00535599 (* 0.0272727 = 0.000146072 loss) | |
| I0525 03:00:04.986301 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00524647 (* 0.0272727 = 0.000143086 loss) | |
| I0525 03:00:04.986311 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00177785 (* 0.0272727 = 4.84869e-05 loss) | |
| I0525 03:00:04.986321 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00240523 (* 0.0272727 = 6.55973e-05 loss) | |
| I0525 03:00:04.986333 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0869565 | |
| I0525 03:00:04.986346 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 03:00:04.986357 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 03:00:04.986369 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0525 03:00:04.986382 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 03:00:04.986392 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 03:00:04.986403 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 03:00:04.986415 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 03:00:04.986428 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 03:00:04.986438 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:00:04.986450 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:00:04.986461 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:00:04.986472 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:00:04.986485 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:00:04.986495 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:00:04.986506 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:00:04.986518 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:00:04.986539 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:00:04.986552 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:00:04.986563 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:00:04.986575 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:00:04.986587 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:00:04.986598 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:00:04.986609 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0525 03:00:04.986620 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.195652 | |
| I0525 03:00:04.986634 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.90268 (* 1 = 2.90268 loss) | |
| I0525 03:00:04.986649 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.854645 (* 1 = 0.854645 loss) | |
| I0525 03:00:04.986663 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.11388 (* 0.0909091 = 0.28308 loss) | |
| I0525 03:00:04.986676 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.04574 (* 0.0909091 = 0.276886 loss) | |
| I0525 03:00:04.986690 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.70287 (* 0.0909091 = 0.245715 loss) | |
| I0525 03:00:04.986703 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.72335 (* 0.0909091 = 0.338486 loss) | |
| I0525 03:00:04.986717 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.31195 (* 0.0909091 = 0.210178 loss) | |
| I0525 03:00:04.986731 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.91392 (* 0.0909091 = 0.173993 loss) | |
| I0525 03:00:04.986744 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.65413 (* 0.0909091 = 0.150376 loss) | |
| I0525 03:00:04.986758 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.674919 (* 0.0909091 = 0.0613563 loss) | |
| I0525 03:00:04.986776 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00161501 (* 0.0909091 = 0.000146819 loss) | |
| I0525 03:00:04.986804 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00139108 (* 0.0909091 = 0.000126461 loss) | |
| I0525 03:00:04.986831 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00119534 (* 0.0909091 = 0.000108667 loss) | |
| I0525 03:00:04.986852 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00134054 (* 0.0909091 = 0.000121868 loss) | |
| I0525 03:00:04.986866 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00124815 (* 0.0909091 = 0.000113468 loss) | |
| I0525 03:00:04.986881 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000974644 (* 0.0909091 = 8.8604e-05 loss) | |
| I0525 03:00:04.986896 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000885804 (* 0.0909091 = 8.05277e-05 loss) | |
| I0525 03:00:04.986909 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000582676 (* 0.0909091 = 5.29705e-05 loss) | |
| I0525 03:00:04.986923 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000812431 (* 0.0909091 = 7.38574e-05 loss) | |
| I0525 03:00:04.986940 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000430967 (* 0.0909091 = 3.91788e-05 loss) | |
| I0525 03:00:04.986955 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000466808 (* 0.0909091 = 4.24371e-05 loss) | |
| I0525 03:00:04.986969 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000470292 (* 0.0909091 = 4.27539e-05 loss) | |
| I0525 03:00:04.986984 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000339967 (* 0.0909091 = 3.09061e-05 loss) | |
| I0525 03:00:04.986997 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000196093 (* 0.0909091 = 1.78267e-05 loss) | |
| I0525 03:00:04.987010 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:00:04.987021 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:00:04.987040 5272 solver.cpp:245] Train net output #149: total_confidence = 4.47047e-05 | |
| I0525 03:00:04.987049 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00067437 | |
| I0525 03:00:04.987067 5272 sgd_solver.cpp:106] Iteration 13000, lr = 0.001 | |
| I0525 03:00:52.282081 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.5046 > 30) by scale factor 0.674088 | |
| I0525 03:03:23.126121 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.2313 > 30) by scale factor 0.960573 | |
| I0525 03:04:00.079668 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.3914 > 30) by scale factor 0.987121 | |
| I0525 03:05:35.481935 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.8578 > 30) by scale factor 0.836637 | |
| I0525 03:06:01.635527 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.9476 > 30) by scale factor 0.715178 | |
| I0525 03:06:13.186782 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.5609 > 30) by scale factor 0.843624 | |
| I0525 03:06:29.743352 5272 solver.cpp:229] Iteration 13500, loss = 10.1399 | |
| I0525 03:06:29.743428 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0512821 | |
| I0525 03:06:29.743448 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 03:06:29.743461 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 03:06:29.743474 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 03:06:29.743487 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 03:06:29.743499 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.5 | |
| I0525 03:06:29.743512 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 03:06:29.743525 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 03:06:29.743537 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 03:06:29.743549 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:06:29.743562 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:06:29.743576 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:06:29.743587 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:06:29.743599 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:06:29.743613 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:06:29.743623 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:06:29.743635 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:06:29.743648 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:06:29.743659 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:06:29.743671 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:06:29.743683 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:06:29.743695 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:06:29.743706 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:06:29.743718 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.789773 | |
| I0525 03:06:29.743731 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.153846 | |
| I0525 03:06:29.743751 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.31886 (* 0.3 = 0.995658 loss) | |
| I0525 03:06:29.743765 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.7997 (* 0.3 = 0.23991 loss) | |
| I0525 03:06:29.743779 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.25373 (* 0.0272727 = 0.088738 loss) | |
| I0525 03:06:29.743793 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.12515 (* 0.0272727 = 0.0852313 loss) | |
| I0525 03:06:29.743808 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.80012 (* 0.0272727 = 0.10364 loss) | |
| I0525 03:06:29.743823 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.39482 (* 0.0272727 = 0.0925861 loss) | |
| I0525 03:06:29.743836 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 1.88191 (* 0.0272727 = 0.0513248 loss) | |
| I0525 03:06:29.743850 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.94251 (* 0.0272727 = 0.0529775 loss) | |
| I0525 03:06:29.743865 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.04494 (* 0.0272727 = 0.0284982 loss) | |
| I0525 03:06:29.743878 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.119656 (* 0.0272727 = 0.00326336 loss) | |
| I0525 03:06:29.743893 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0388553 (* 0.0272727 = 0.00105969 loss) | |
| I0525 03:06:29.743907 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.022958 (* 0.0272727 = 0.000626126 loss) | |
| I0525 03:06:29.743921 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0172133 (* 0.0272727 = 0.000469453 loss) | |
| I0525 03:06:29.743935 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0068379 (* 0.0272727 = 0.000186488 loss) | |
| I0525 03:06:29.743993 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00559391 (* 0.0272727 = 0.000152561 loss) | |
| I0525 03:06:29.744009 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00366881 (* 0.0272727 = 0.000100058 loss) | |
| I0525 03:06:29.744022 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00343014 (* 0.0272727 = 9.35493e-05 loss) | |
| I0525 03:06:29.744036 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00253565 (* 0.0272727 = 6.91541e-05 loss) | |
| I0525 03:06:29.744050 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00152215 (* 0.0272727 = 4.15133e-05 loss) | |
| I0525 03:06:29.744065 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0011334 (* 0.0272727 = 3.0911e-05 loss) | |
| I0525 03:06:29.744078 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00078968 (* 0.0272727 = 2.15367e-05 loss) | |
| I0525 03:06:29.744092 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00167102 (* 0.0272727 = 4.55733e-05 loss) | |
| I0525 03:06:29.744107 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.000710153 (* 0.0272727 = 1.93678e-05 loss) | |
| I0525 03:06:29.744120 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000525491 (* 0.0272727 = 1.43316e-05 loss) | |
| I0525 03:06:29.744133 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.025641 | |
| I0525 03:06:29.744145 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 03:06:29.744158 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 03:06:29.744169 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 03:06:29.744181 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0525 03:06:29.744192 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.5 | |
| I0525 03:06:29.744205 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 03:06:29.744217 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 03:06:29.744230 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 03:06:29.744241 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:06:29.744252 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:06:29.744264 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:06:29.744276 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:06:29.744287 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:06:29.744299 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:06:29.744310 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:06:29.744323 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:06:29.744333 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:06:29.744344 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:06:29.744356 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:06:29.744367 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:06:29.744379 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:06:29.744391 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:06:29.744402 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.772727 | |
| I0525 03:06:29.744415 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.230769 | |
| I0525 03:06:29.744428 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.39419 (* 0.3 = 1.01826 loss) | |
| I0525 03:06:29.744441 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.851331 (* 0.3 = 0.255399 loss) | |
| I0525 03:06:29.744456 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.60404 (* 0.0272727 = 0.098292 loss) | |
| I0525 03:06:29.744469 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.97243 (* 0.0272727 = 0.0810663 loss) | |
| I0525 03:06:29.744494 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.44948 (* 0.0272727 = 0.0940767 loss) | |
| I0525 03:06:29.744508 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.01737 (* 0.0272727 = 0.0822918 loss) | |
| I0525 03:06:29.744523 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 1.56064 (* 0.0272727 = 0.042563 loss) | |
| I0525 03:06:29.744536 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 1.93347 (* 0.0272727 = 0.0527311 loss) | |
| I0525 03:06:29.744550 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 0.97303 (* 0.0272727 = 0.0265372 loss) | |
| I0525 03:06:29.744565 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.161033 (* 0.0272727 = 0.00439181 loss) | |
| I0525 03:06:29.744578 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0764242 (* 0.0272727 = 0.0020843 loss) | |
| I0525 03:06:29.744593 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0229605 (* 0.0272727 = 0.000626195 loss) | |
| I0525 03:06:29.744606 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0196934 (* 0.0272727 = 0.000537094 loss) | |
| I0525 03:06:29.744621 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0112483 (* 0.0272727 = 0.000306771 loss) | |
| I0525 03:06:29.744634 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00668883 (* 0.0272727 = 0.000182423 loss) | |
| I0525 03:06:29.744648 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0084138 (* 0.0272727 = 0.000229467 loss) | |
| I0525 03:06:29.744663 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00506988 (* 0.0272727 = 0.00013827 loss) | |
| I0525 03:06:29.744676 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00436511 (* 0.0272727 = 0.000119049 loss) | |
| I0525 03:06:29.744690 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00406082 (* 0.0272727 = 0.00011075 loss) | |
| I0525 03:06:29.744704 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00206583 (* 0.0272727 = 5.63408e-05 loss) | |
| I0525 03:06:29.744719 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0012821 (* 0.0272727 = 3.49664e-05 loss) | |
| I0525 03:06:29.744732 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00117685 (* 0.0272727 = 3.20958e-05 loss) | |
| I0525 03:06:29.744746 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00127415 (* 0.0272727 = 3.47495e-05 loss) | |
| I0525 03:06:29.744760 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00113219 (* 0.0272727 = 3.08779e-05 loss) | |
| I0525 03:06:29.744772 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.102564 | |
| I0525 03:06:29.744784 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 03:06:29.744799 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 03:06:29.744812 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 03:06:29.744824 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 03:06:29.744837 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0525 03:06:29.744848 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 03:06:29.744861 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 03:06:29.744874 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 03:06:29.744885 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:06:29.744896 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:06:29.744909 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:06:29.744920 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:06:29.744933 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:06:29.744942 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:06:29.744954 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:06:29.744976 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:06:29.744988 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:06:29.745000 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:06:29.745012 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:06:29.745028 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:06:29.745039 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:06:29.745051 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:06:29.745064 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.801136 | |
| I0525 03:06:29.745075 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.307692 | |
| I0525 03:06:29.745090 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.0596 (* 1 = 3.0596 loss) | |
| I0525 03:06:29.745103 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.742618 (* 1 = 0.742618 loss) | |
| I0525 03:06:29.745129 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.24854 (* 0.0909091 = 0.295322 loss) | |
| I0525 03:06:29.745147 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.96913 (* 0.0909091 = 0.269921 loss) | |
| I0525 03:06:29.745162 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.21857 (* 0.0909091 = 0.292597 loss) | |
| I0525 03:06:29.745175 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.81778 (* 0.0909091 = 0.256162 loss) | |
| I0525 03:06:29.745189 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 1.36325 (* 0.0909091 = 0.123932 loss) | |
| I0525 03:06:29.745203 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.47907 (* 0.0909091 = 0.134461 loss) | |
| I0525 03:06:29.745218 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.77484 (* 0.0909091 = 0.07044 loss) | |
| I0525 03:06:29.745230 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.12039 (* 0.0909091 = 0.0109446 loss) | |
| I0525 03:06:29.745245 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0433997 (* 0.0909091 = 0.00394543 loss) | |
| I0525 03:06:29.745260 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0172939 (* 0.0909091 = 0.00157217 loss) | |
| I0525 03:06:29.745273 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00502935 (* 0.0909091 = 0.000457214 loss) | |
| I0525 03:06:29.745287 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00371122 (* 0.0909091 = 0.000337383 loss) | |
| I0525 03:06:29.745301 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00483404 (* 0.0909091 = 0.000439459 loss) | |
| I0525 03:06:29.745314 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00309594 (* 0.0909091 = 0.000281449 loss) | |
| I0525 03:06:29.745328 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00215051 (* 0.0909091 = 0.000195501 loss) | |
| I0525 03:06:29.745342 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00257554 (* 0.0909091 = 0.00023414 loss) | |
| I0525 03:06:29.745357 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00275208 (* 0.0909091 = 0.000250189 loss) | |
| I0525 03:06:29.745369 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00177781 (* 0.0909091 = 0.000161619 loss) | |
| I0525 03:06:29.745383 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00148582 (* 0.0909091 = 0.000135074 loss) | |
| I0525 03:06:29.745398 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00131686 (* 0.0909091 = 0.000119714 loss) | |
| I0525 03:06:29.745411 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000900455 (* 0.0909091 = 8.18596e-05 loss) | |
| I0525 03:06:29.745425 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00069871 (* 0.0909091 = 6.35191e-05 loss) | |
| I0525 03:06:29.745439 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:06:29.745450 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:06:29.745472 5272 solver.cpp:245] Train net output #149: total_confidence = 0.00010129 | |
| I0525 03:06:29.745486 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000350062 | |
| I0525 03:06:29.745499 5272 sgd_solver.cpp:106] Iteration 13500, lr = 0.001 | |
| I0525 03:08:14.797674 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.9022 > 30) by scale factor 0.859545 | |
| I0525 03:09:30.191488 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.4005 > 30) by scale factor 0.872079 | |
| I0525 03:11:31.725409 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 45.2166 > 30) by scale factor 0.663474 | |
| I0525 03:12:54.462980 5272 solver.cpp:229] Iteration 14000, loss = 10.1608 | |
| I0525 03:12:54.463075 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0943396 | |
| I0525 03:12:54.463095 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 03:12:54.463111 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 03:12:54.463124 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 03:12:54.463137 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 03:12:54.463150 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 03:12:54.463163 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 03:12:54.463176 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 03:12:54.463188 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 03:12:54.463201 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 03:12:54.463213 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 03:12:54.463227 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 03:12:54.463239 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 03:12:54.463251 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 03:12:54.463264 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 03:12:54.463276 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 03:12:54.463289 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 03:12:54.463300 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:12:54.463312 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:12:54.463325 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:12:54.463336 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:12:54.463348 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:12:54.463361 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:12:54.463372 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.710227 | |
| I0525 03:12:54.463385 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.283019 | |
| I0525 03:12:54.463402 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.03152 (* 0.3 = 0.909457 loss) | |
| I0525 03:12:54.463416 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.06279 (* 0.3 = 0.318837 loss) | |
| I0525 03:12:54.463430 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.74712 (* 0.0272727 = 0.0749213 loss) | |
| I0525 03:12:54.463444 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.18125 (* 0.0272727 = 0.0867614 loss) | |
| I0525 03:12:54.463459 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.38212 (* 0.0272727 = 0.0922397 loss) | |
| I0525 03:12:54.463472 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.21076 (* 0.0272727 = 0.0875663 loss) | |
| I0525 03:12:54.463486 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.21107 (* 0.0272727 = 0.0875746 loss) | |
| I0525 03:12:54.463500 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.11388 (* 0.0272727 = 0.0576512 loss) | |
| I0525 03:12:54.463515 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 0.859994 (* 0.0272727 = 0.0234544 loss) | |
| I0525 03:12:54.463528 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.21496 (* 0.0272727 = 0.0331353 loss) | |
| I0525 03:12:54.463546 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.327197 (* 0.0272727 = 0.00892355 loss) | |
| I0525 03:12:54.463559 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.660401 (* 0.0272727 = 0.0180109 loss) | |
| I0525 03:12:54.463573 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.42954 (* 0.0272727 = 0.0117147 loss) | |
| I0525 03:12:54.463587 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.400203 (* 0.0272727 = 0.0109146 loss) | |
| I0525 03:12:54.463620 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.46168 (* 0.0272727 = 0.0125913 loss) | |
| I0525 03:12:54.463636 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.357086 (* 0.0272727 = 0.00973871 loss) | |
| I0525 03:12:54.463651 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.495131 (* 0.0272727 = 0.0135036 loss) | |
| I0525 03:12:54.463665 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.569613 (* 0.0272727 = 0.0155349 loss) | |
| I0525 03:12:54.463680 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.016396 (* 0.0272727 = 0.000447164 loss) | |
| I0525 03:12:54.463693 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00935248 (* 0.0272727 = 0.000255068 loss) | |
| I0525 03:12:54.463708 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00772858 (* 0.0272727 = 0.00021078 loss) | |
| I0525 03:12:54.463722 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00954462 (* 0.0272727 = 0.000260308 loss) | |
| I0525 03:12:54.463735 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00426503 (* 0.0272727 = 0.000116319 loss) | |
| I0525 03:12:54.463749 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00330752 (* 0.0272727 = 9.02052e-05 loss) | |
| I0525 03:12:54.463765 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0566038 | |
| I0525 03:12:54.463778 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 03:12:54.463790 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 03:12:54.463803 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 03:12:54.463814 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 03:12:54.463826 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 03:12:54.463838 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 03:12:54.463850 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 03:12:54.463862 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 03:12:54.463874 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:12:54.463886 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 03:12:54.463897 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 03:12:54.463909 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 03:12:54.463922 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 03:12:54.463933 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 03:12:54.463945 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 03:12:54.463958 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 03:12:54.463969 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:12:54.463981 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:12:54.463994 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:12:54.464004 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:12:54.464016 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:12:54.464028 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:12:54.464040 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.693182 | |
| I0525 03:12:54.464051 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.283019 | |
| I0525 03:12:54.464066 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.94768 (* 0.3 = 0.884305 loss) | |
| I0525 03:12:54.464079 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.03152 (* 0.3 = 0.309456 loss) | |
| I0525 03:12:54.464093 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.90623 (* 0.0272727 = 0.0792609 loss) | |
| I0525 03:12:54.464107 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.36424 (* 0.0272727 = 0.091752 loss) | |
| I0525 03:12:54.464131 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.54173 (* 0.0272727 = 0.0965927 loss) | |
| I0525 03:12:54.464146 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.32354 (* 0.0272727 = 0.090642 loss) | |
| I0525 03:12:54.464164 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.92551 (* 0.0272727 = 0.0797867 loss) | |
| I0525 03:12:54.464177 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.10071 (* 0.0272727 = 0.057292 loss) | |
| I0525 03:12:54.464191 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.10607 (* 0.0272727 = 0.0301655 loss) | |
| I0525 03:12:54.464205 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.13735 (* 0.0272727 = 0.0310185 loss) | |
| I0525 03:12:54.464220 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.267548 (* 0.0272727 = 0.00729677 loss) | |
| I0525 03:12:54.464233 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.596094 (* 0.0272727 = 0.0162571 loss) | |
| I0525 03:12:54.464248 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.623699 (* 0.0272727 = 0.01701 loss) | |
| I0525 03:12:54.464262 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.432461 (* 0.0272727 = 0.0117944 loss) | |
| I0525 03:12:54.464275 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.353406 (* 0.0272727 = 0.00963835 loss) | |
| I0525 03:12:54.464289 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.550262 (* 0.0272727 = 0.0150071 loss) | |
| I0525 03:12:54.464303 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.604975 (* 0.0272727 = 0.0164993 loss) | |
| I0525 03:12:54.464318 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.59506 (* 0.0272727 = 0.0162289 loss) | |
| I0525 03:12:54.464332 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0262886 (* 0.0272727 = 0.000716963 loss) | |
| I0525 03:12:54.464345 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0137888 (* 0.0272727 = 0.000376059 loss) | |
| I0525 03:12:54.464360 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.013693 (* 0.0272727 = 0.000373445 loss) | |
| I0525 03:12:54.464375 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0141221 (* 0.0272727 = 0.000385148 loss) | |
| I0525 03:12:54.464388 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00419674 (* 0.0272727 = 0.000114457 loss) | |
| I0525 03:12:54.464402 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00959846 (* 0.0272727 = 0.000261776 loss) | |
| I0525 03:12:54.464414 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0754717 | |
| I0525 03:12:54.464426 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 03:12:54.464438 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0525 03:12:54.464452 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 03:12:54.464463 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 03:12:54.464474 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 03:12:54.464486 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 03:12:54.464498 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 03:12:54.464510 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 03:12:54.464521 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 03:12:54.464534 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 03:12:54.464545 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 03:12:54.464557 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 03:12:54.464570 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 03:12:54.464581 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 03:12:54.464592 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 03:12:54.464614 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 03:12:54.464629 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:12:54.464637 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:12:54.464650 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:12:54.464663 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:12:54.464673 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:12:54.464685 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:12:54.464697 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.710227 | |
| I0525 03:12:54.464709 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.283019 | |
| I0525 03:12:54.464723 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.9011 (* 1 = 2.9011 loss) | |
| I0525 03:12:54.464737 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.993806 (* 1 = 0.993806 loss) | |
| I0525 03:12:54.464751 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.50495 (* 0.0909091 = 0.227723 loss) | |
| I0525 03:12:54.464766 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.10065 (* 0.0909091 = 0.281877 loss) | |
| I0525 03:12:54.464779 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.46901 (* 0.0909091 = 0.315364 loss) | |
| I0525 03:12:54.464793 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.64199 (* 0.0909091 = 0.240181 loss) | |
| I0525 03:12:54.464807 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.66086 (* 0.0909091 = 0.241896 loss) | |
| I0525 03:12:54.464824 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.10047 (* 0.0909091 = 0.190951 loss) | |
| I0525 03:12:54.464838 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.927365 (* 0.0909091 = 0.0843059 loss) | |
| I0525 03:12:54.464853 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.09841 (* 0.0909091 = 0.0998558 loss) | |
| I0525 03:12:54.464866 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.257948 (* 0.0909091 = 0.0234499 loss) | |
| I0525 03:12:54.464880 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.72168 (* 0.0909091 = 0.0656073 loss) | |
| I0525 03:12:54.464895 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.555555 (* 0.0909091 = 0.050505 loss) | |
| I0525 03:12:54.464908 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.63542 (* 0.0909091 = 0.0577655 loss) | |
| I0525 03:12:54.464922 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.456553 (* 0.0909091 = 0.0415048 loss) | |
| I0525 03:12:54.464936 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.49912 (* 0.0909091 = 0.0453746 loss) | |
| I0525 03:12:54.464951 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.447527 (* 0.0909091 = 0.0406843 loss) | |
| I0525 03:12:54.464964 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.54897 (* 0.0909091 = 0.0499064 loss) | |
| I0525 03:12:54.464978 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00387471 (* 0.0909091 = 0.000352246 loss) | |
| I0525 03:12:54.464993 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00247859 (* 0.0909091 = 0.000225326 loss) | |
| I0525 03:12:54.465008 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00288135 (* 0.0909091 = 0.000261941 loss) | |
| I0525 03:12:54.465020 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00199457 (* 0.0909091 = 0.000181324 loss) | |
| I0525 03:12:54.465034 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000966903 (* 0.0909091 = 8.79003e-05 loss) | |
| I0525 03:12:54.465049 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000568925 (* 0.0909091 = 5.17204e-05 loss) | |
| I0525 03:12:54.465061 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:12:54.465072 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:12:54.465085 5272 solver.cpp:245] Train net output #149: total_confidence = 3.17824e-05 | |
| I0525 03:12:54.465106 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000568825 | |
| I0525 03:12:54.465137 5272 sgd_solver.cpp:106] Iteration 14000, lr = 0.001 | |
| I0525 03:13:24.824494 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.6999 > 30) by scale factor 0.817442 | |
| I0525 03:14:09.448601 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.246 > 30) by scale factor 0.693705 | |
| I0525 03:17:43.520668 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.9116 > 30) by scale factor 0.97051 | |
| I0525 03:18:25.106947 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.9359 > 30) by scale factor 0.884021 | |
| I0525 03:19:19.519412 5272 solver.cpp:229] Iteration 14500, loss = 10.0966 | |
| I0525 03:19:19.519532 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0652174 | |
| I0525 03:19:19.519572 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 03:19:19.519588 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 03:19:19.519601 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 03:19:19.519614 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 03:19:19.519626 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 03:19:19.519639 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 03:19:19.519652 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 03:19:19.519665 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 03:19:19.519677 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:19:19.519690 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:19:19.519702 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:19:19.519714 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:19:19.519726 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:19:19.519740 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:19:19.519753 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:19:19.519765 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:19:19.519778 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:19:19.519789 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:19:19.519801 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:19:19.519814 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:19:19.519824 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:19:19.519836 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:19:19.519848 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0525 03:19:19.519860 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.217391 | |
| I0525 03:19:19.519877 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.99777 (* 0.3 = 0.899331 loss) | |
| I0525 03:19:19.519891 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.953346 (* 0.3 = 0.286004 loss) | |
| I0525 03:19:19.519906 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.38889 (* 0.0272727 = 0.0924243 loss) | |
| I0525 03:19:19.519920 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.35098 (* 0.0272727 = 0.0913905 loss) | |
| I0525 03:19:19.519934 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.41937 (* 0.0272727 = 0.0932556 loss) | |
| I0525 03:19:19.519948 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.25054 (* 0.0272727 = 0.0886512 loss) | |
| I0525 03:19:19.519961 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.39881 (* 0.0272727 = 0.0654221 loss) | |
| I0525 03:19:19.519975 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.04982 (* 0.0272727 = 0.0831769 loss) | |
| I0525 03:19:19.519989 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.18057 (* 0.0272727 = 0.0321974 loss) | |
| I0525 03:19:19.520004 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.00907 (* 0.0272727 = 0.0275201 loss) | |
| I0525 03:19:19.520017 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0125308 (* 0.0272727 = 0.000341749 loss) | |
| I0525 03:19:19.520031 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0220301 (* 0.0272727 = 0.000600821 loss) | |
| I0525 03:19:19.520046 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0112052 (* 0.0272727 = 0.000305598 loss) | |
| I0525 03:19:19.520059 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0111337 (* 0.0272727 = 0.000303647 loss) | |
| I0525 03:19:19.520073 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0162398 (* 0.0272727 = 0.000442905 loss) | |
| I0525 03:19:19.520112 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00860229 (* 0.0272727 = 0.000234608 loss) | |
| I0525 03:19:19.520128 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00512887 (* 0.0272727 = 0.000139878 loss) | |
| I0525 03:19:19.520143 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0106722 (* 0.0272727 = 0.00029106 loss) | |
| I0525 03:19:19.520156 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00517896 (* 0.0272727 = 0.000141244 loss) | |
| I0525 03:19:19.520170 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0120481 (* 0.0272727 = 0.000328584 loss) | |
| I0525 03:19:19.520184 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0081009 (* 0.0272727 = 0.000220934 loss) | |
| I0525 03:19:19.520198 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00486691 (* 0.0272727 = 0.000132734 loss) | |
| I0525 03:19:19.520212 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00991399 (* 0.0272727 = 0.000270381 loss) | |
| I0525 03:19:19.520226 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00700163 (* 0.0272727 = 0.000190953 loss) | |
| I0525 03:19:19.520239 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0652174 | |
| I0525 03:19:19.520251 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 03:19:19.520262 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 03:19:19.520274 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 03:19:19.520287 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 03:19:19.520298 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 03:19:19.520310 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 03:19:19.520323 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 03:19:19.520334 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 03:19:19.520345 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:19:19.520357 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:19:19.520369 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:19:19.520380 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:19:19.520391 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:19:19.520403 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:19:19.520414 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:19:19.520426 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:19:19.520438 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:19:19.520449 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:19:19.520462 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:19:19.520473 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:19:19.520484 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:19:19.520496 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:19:19.520508 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0525 03:19:19.520519 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.282609 | |
| I0525 03:19:19.520534 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.85531 (* 0.3 = 0.856593 loss) | |
| I0525 03:19:19.520547 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.936045 (* 0.3 = 0.280814 loss) | |
| I0525 03:19:19.520561 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.01637 (* 0.0272727 = 0.0822647 loss) | |
| I0525 03:19:19.520575 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.46897 (* 0.0272727 = 0.0673355 loss) | |
| I0525 03:19:19.520599 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.47579 (* 0.0272727 = 0.0947942 loss) | |
| I0525 03:19:19.520614 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.17222 (* 0.0272727 = 0.086515 loss) | |
| I0525 03:19:19.520628 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.71052 (* 0.0272727 = 0.0739233 loss) | |
| I0525 03:19:19.520642 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.84651 (* 0.0272727 = 0.077632 loss) | |
| I0525 03:19:19.520655 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.20185 (* 0.0272727 = 0.0327776 loss) | |
| I0525 03:19:19.520669 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.703402 (* 0.0272727 = 0.0191837 loss) | |
| I0525 03:19:19.520684 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0106011 (* 0.0272727 = 0.000289121 loss) | |
| I0525 03:19:19.520697 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00664626 (* 0.0272727 = 0.000181262 loss) | |
| I0525 03:19:19.520711 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00321078 (* 0.0272727 = 8.75668e-05 loss) | |
| I0525 03:19:19.520726 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00312367 (* 0.0272727 = 8.5191e-05 loss) | |
| I0525 03:19:19.520740 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00236252 (* 0.0272727 = 6.44324e-05 loss) | |
| I0525 03:19:19.520755 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00188441 (* 0.0272727 = 5.1393e-05 loss) | |
| I0525 03:19:19.520768 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00349312 (* 0.0272727 = 9.5267e-05 loss) | |
| I0525 03:19:19.520782 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00231733 (* 0.0272727 = 6.32e-05 loss) | |
| I0525 03:19:19.520799 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00142812 (* 0.0272727 = 3.89487e-05 loss) | |
| I0525 03:19:19.520814 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00283683 (* 0.0272727 = 7.73681e-05 loss) | |
| I0525 03:19:19.520828 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00179865 (* 0.0272727 = 4.9054e-05 loss) | |
| I0525 03:19:19.520843 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00261521 (* 0.0272727 = 7.13239e-05 loss) | |
| I0525 03:19:19.520856 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00348873 (* 0.0272727 = 9.51473e-05 loss) | |
| I0525 03:19:19.520870 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00274323 (* 0.0272727 = 7.48153e-05 loss) | |
| I0525 03:19:19.520884 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.173913 | |
| I0525 03:19:19.520895 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 03:19:19.520907 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 03:19:19.520920 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.375 | |
| I0525 03:19:19.520931 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 03:19:19.520943 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 03:19:19.520956 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 03:19:19.520967 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 03:19:19.520979 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 03:19:19.520990 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:19:19.521003 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:19:19.521013 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:19:19.521025 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:19:19.521037 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:19:19.521049 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:19:19.521060 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:19:19.521072 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:19:19.521093 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:19:19.521106 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:19:19.521136 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:19:19.521152 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:19:19.521162 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:19:19.521174 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:19:19.521186 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.778409 | |
| I0525 03:19:19.521198 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.391304 | |
| I0525 03:19:19.521209 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.69064 (* 1 = 2.69064 loss) | |
| I0525 03:19:19.521219 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.800994 (* 1 = 0.800994 loss) | |
| I0525 03:19:19.521234 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.33833 (* 0.0909091 = 0.212576 loss) | |
| I0525 03:19:19.521248 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.65436 (* 0.0909091 = 0.241306 loss) | |
| I0525 03:19:19.521262 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.00796 (* 0.0909091 = 0.273451 loss) | |
| I0525 03:19:19.521276 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.33343 (* 0.0909091 = 0.303039 loss) | |
| I0525 03:19:19.521289 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.31031 (* 0.0909091 = 0.210029 loss) | |
| I0525 03:19:19.521303 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.27319 (* 0.0909091 = 0.206654 loss) | |
| I0525 03:19:19.521317 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.987878 (* 0.0909091 = 0.0898071 loss) | |
| I0525 03:19:19.521330 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.421453 (* 0.0909091 = 0.0383139 loss) | |
| I0525 03:19:19.521344 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00227712 (* 0.0909091 = 0.000207011 loss) | |
| I0525 03:19:19.521358 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00146184 (* 0.0909091 = 0.000132895 loss) | |
| I0525 03:19:19.521373 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00150392 (* 0.0909091 = 0.00013672 loss) | |
| I0525 03:19:19.521386 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.000993792 (* 0.0909091 = 9.03447e-05 loss) | |
| I0525 03:19:19.521400 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00130384 (* 0.0909091 = 0.000118531 loss) | |
| I0525 03:19:19.521414 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000975736 (* 0.0909091 = 8.87033e-05 loss) | |
| I0525 03:19:19.521428 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00102532 (* 0.0909091 = 9.32111e-05 loss) | |
| I0525 03:19:19.521441 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000838571 (* 0.0909091 = 7.62337e-05 loss) | |
| I0525 03:19:19.521456 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000741782 (* 0.0909091 = 6.74347e-05 loss) | |
| I0525 03:19:19.521468 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000543347 (* 0.0909091 = 4.93951e-05 loss) | |
| I0525 03:19:19.521482 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00108501 (* 0.0909091 = 9.86376e-05 loss) | |
| I0525 03:19:19.521497 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00103212 (* 0.0909091 = 9.38293e-05 loss) | |
| I0525 03:19:19.521510 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.0015249 (* 0.0909091 = 0.000138627 loss) | |
| I0525 03:19:19.521524 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000969557 (* 0.0909091 = 8.81415e-05 loss) | |
| I0525 03:19:19.521536 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:19:19.521548 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:19:19.521570 5272 solver.cpp:245] Train net output #149: total_confidence = 1.38009e-05 | |
| I0525 03:19:19.521584 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000273296 | |
| I0525 03:19:19.521596 5272 sgd_solver.cpp:106] Iteration 14500, lr = 0.001 | |
| I0525 03:23:33.840852 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.7756 > 30) by scale factor 0.915315 | |
| I0525 03:25:43.898362 5272 solver.cpp:338] Iteration 15000, Testing net (#0) | |
| I0525 03:26:41.757014 5272 solver.cpp:393] Test loss: 9.34947 | |
| I0525 03:26:41.757114 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0662929 | |
| I0525 03:26:41.757148 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.126 | |
| I0525 03:26:41.757163 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.097 | |
| I0525 03:26:41.757175 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.088 | |
| I0525 03:26:41.757187 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.168 | |
| I0525 03:26:41.757200 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.319 | |
| I0525 03:26:41.757212 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.47 | |
| I0525 03:26:41.757225 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.74 | |
| I0525 03:26:41.757236 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.919 | |
| I0525 03:26:41.757248 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.98 | |
| I0525 03:26:41.757261 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.994 | |
| I0525 03:26:41.757272 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0525 03:26:41.757283 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0525 03:26:41.757295 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0525 03:26:41.757307 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0525 03:26:41.757318 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0525 03:26:41.757329 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0525 03:26:41.757340 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0525 03:26:41.757351 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0525 03:26:41.757362 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0525 03:26:41.757375 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0525 03:26:41.757385 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0525 03:26:41.757396 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0525 03:26:41.757407 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.76591 | |
| I0525 03:26:41.757421 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.21944 | |
| I0525 03:26:41.757436 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.56259 (* 0.3 = 1.06878 loss) | |
| I0525 03:26:41.757450 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.921492 (* 0.3 = 0.276448 loss) | |
| I0525 03:26:41.757465 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 2.95648 (* 0.0272727 = 0.0806314 loss) | |
| I0525 03:26:41.757478 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.18494 (* 0.0272727 = 0.0868619 loss) | |
| I0525 03:26:41.757493 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.26209 (* 0.0272727 = 0.0889662 loss) | |
| I0525 03:26:41.757505 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.12925 (* 0.0272727 = 0.0853431 loss) | |
| I0525 03:26:41.757519 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 2.64207 (* 0.0272727 = 0.0720564 loss) | |
| I0525 03:26:41.757532 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.18369 (* 0.0272727 = 0.0595551 loss) | |
| I0525 03:26:41.757545 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.2699 (* 0.0272727 = 0.0346335 loss) | |
| I0525 03:26:41.757560 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.457305 (* 0.0272727 = 0.0124719 loss) | |
| I0525 03:26:41.757573 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.121033 (* 0.0272727 = 0.00330091 loss) | |
| I0525 03:26:41.757591 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0675041 (* 0.0272727 = 0.00184102 loss) | |
| I0525 03:26:41.757606 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.0309848 (* 0.0272727 = 0.000845039 loss) | |
| I0525 03:26:41.757619 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0234151 (* 0.0272727 = 0.000638593 loss) | |
| I0525 03:26:41.757633 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0188042 (* 0.0272727 = 0.000512841 loss) | |
| I0525 03:26:41.757671 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0153167 (* 0.0272727 = 0.000417729 loss) | |
| I0525 03:26:41.757686 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.0112174 (* 0.0272727 = 0.000305928 loss) | |
| I0525 03:26:41.757700 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.00827308 (* 0.0272727 = 0.000225629 loss) | |
| I0525 03:26:41.757714 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.00556261 (* 0.0272727 = 0.000151708 loss) | |
| I0525 03:26:41.757727 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00428337 (* 0.0272727 = 0.000116819 loss) | |
| I0525 03:26:41.757741 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0045138 (* 0.0272727 = 0.000123104 loss) | |
| I0525 03:26:41.757755 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00391397 (* 0.0272727 = 0.000106745 loss) | |
| I0525 03:26:41.757769 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00371399 (* 0.0272727 = 0.000101291 loss) | |
| I0525 03:26:41.757783 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00343956 (* 0.0272727 = 9.38063e-05 loss) | |
| I0525 03:26:41.757796 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.062804 | |
| I0525 03:26:41.757807 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.119 | |
| I0525 03:26:41.757819 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.098 | |
| I0525 03:26:41.757832 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.083 | |
| I0525 03:26:41.757843 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.161 | |
| I0525 03:26:41.757853 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.33 | |
| I0525 03:26:41.757865 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.471 | |
| I0525 03:26:41.757876 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.741 | |
| I0525 03:26:41.757887 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.919 | |
| I0525 03:26:41.757899 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.982 | |
| I0525 03:26:41.757910 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.994 | |
| I0525 03:26:41.757921 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0525 03:26:41.757936 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0525 03:26:41.757947 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0525 03:26:41.757958 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0525 03:26:41.757969 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0525 03:26:41.757980 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0525 03:26:41.757992 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0525 03:26:41.758002 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0525 03:26:41.758013 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0525 03:26:41.758023 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0525 03:26:41.758034 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0525 03:26:41.758045 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0525 03:26:41.758056 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.765137 | |
| I0525 03:26:41.758067 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.231335 | |
| I0525 03:26:41.758081 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.54987 (* 0.3 = 1.06496 loss) | |
| I0525 03:26:41.758095 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.921071 (* 0.3 = 0.276321 loss) | |
| I0525 03:26:41.758108 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 2.92427 (* 0.0272727 = 0.0797527 loss) | |
| I0525 03:26:41.758121 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.13984 (* 0.0272727 = 0.0856321 loss) | |
| I0525 03:26:41.758134 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.22669 (* 0.0272727 = 0.0880007 loss) | |
| I0525 03:26:41.758159 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 3.09169 (* 0.0272727 = 0.0843189 loss) | |
| I0525 03:26:41.758174 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 2.60782 (* 0.0272727 = 0.0711223 loss) | |
| I0525 03:26:41.758188 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.15695 (* 0.0272727 = 0.0588259 loss) | |
| I0525 03:26:41.758200 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.24577 (* 0.0272727 = 0.0339755 loss) | |
| I0525 03:26:41.758213 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.445389 (* 0.0272727 = 0.012147 loss) | |
| I0525 03:26:41.758227 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.108186 (* 0.0272727 = 0.00295052 loss) | |
| I0525 03:26:41.758241 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0548634 (* 0.0272727 = 0.00149627 loss) | |
| I0525 03:26:41.758255 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.0180264 (* 0.0272727 = 0.00049163 loss) | |
| I0525 03:26:41.758268 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.0124881 (* 0.0272727 = 0.000340585 loss) | |
| I0525 03:26:41.758281 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.0100058 (* 0.0272727 = 0.000272884 loss) | |
| I0525 03:26:41.758294 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.00807055 (* 0.0272727 = 0.000220106 loss) | |
| I0525 03:26:41.758308 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00670154 (* 0.0272727 = 0.000182769 loss) | |
| I0525 03:26:41.758322 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00492736 (* 0.0272727 = 0.000134383 loss) | |
| I0525 03:26:41.758335 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.00320336 (* 0.0272727 = 8.73644e-05 loss) | |
| I0525 03:26:41.758349 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00302355 (* 0.0272727 = 8.24604e-05 loss) | |
| I0525 03:26:41.758363 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.00286073 (* 0.0272727 = 7.80198e-05 loss) | |
| I0525 03:26:41.758376 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00245152 (* 0.0272727 = 6.68596e-05 loss) | |
| I0525 03:26:41.758390 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00218105 (* 0.0272727 = 5.9483e-05 loss) | |
| I0525 03:26:41.758400 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00229509 (* 0.0272727 = 6.25934e-05 loss) | |
| I0525 03:26:41.758407 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0773798 | |
| I0525 03:26:41.758420 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.123 | |
| I0525 03:26:41.758432 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.094 | |
| I0525 03:26:41.758445 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.08 | |
| I0525 03:26:41.758456 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.169 | |
| I0525 03:26:41.758467 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.326 | |
| I0525 03:26:41.758478 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.478 | |
| I0525 03:26:41.758489 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.74 | |
| I0525 03:26:41.758502 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.918 | |
| I0525 03:26:41.758512 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.979 | |
| I0525 03:26:41.758523 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.992 | |
| I0525 03:26:41.758535 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 1 | |
| I0525 03:26:41.758546 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 1 | |
| I0525 03:26:41.758558 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0525 03:26:41.758569 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0525 03:26:41.758579 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0525 03:26:41.758589 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0525 03:26:41.758600 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0525 03:26:41.758621 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0525 03:26:41.758636 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0525 03:26:41.758648 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0525 03:26:41.758659 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0525 03:26:41.758671 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0525 03:26:41.758682 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.766592 | |
| I0525 03:26:41.758693 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.245081 | |
| I0525 03:26:41.758707 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.12194 (* 1 = 3.12194 loss) | |
| I0525 03:26:41.758720 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.832668 (* 1 = 0.832668 loss) | |
| I0525 03:26:41.758733 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 2.77176 (* 0.0909091 = 0.251978 loss) | |
| I0525 03:26:41.758747 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 3.00864 (* 0.0909091 = 0.273512 loss) | |
| I0525 03:26:41.758760 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.09214 (* 0.0909091 = 0.281104 loss) | |
| I0525 03:26:41.758774 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 2.95636 (* 0.0909091 = 0.26876 loss) | |
| I0525 03:26:41.758786 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.51062 (* 0.0909091 = 0.228238 loss) | |
| I0525 03:26:41.758800 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.04374 (* 0.0909091 = 0.185795 loss) | |
| I0525 03:26:41.758812 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.17651 (* 0.0909091 = 0.106955 loss) | |
| I0525 03:26:41.758826 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.427845 (* 0.0909091 = 0.038895 loss) | |
| I0525 03:26:41.758839 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.107273 (* 0.0909091 = 0.00975207 loss) | |
| I0525 03:26:41.758853 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0562661 (* 0.0909091 = 0.0051151 loss) | |
| I0525 03:26:41.758867 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.0227568 (* 0.0909091 = 0.0020688 loss) | |
| I0525 03:26:41.758880 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0180259 (* 0.0909091 = 0.00163872 loss) | |
| I0525 03:26:41.758893 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.0141487 (* 0.0909091 = 0.00128625 loss) | |
| I0525 03:26:41.758908 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.0125856 (* 0.0909091 = 0.00114414 loss) | |
| I0525 03:26:41.758920 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.00979006 (* 0.0909091 = 0.000890005 loss) | |
| I0525 03:26:41.758934 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00596141 (* 0.0909091 = 0.000541946 loss) | |
| I0525 03:26:41.758947 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00309804 (* 0.0909091 = 0.00028164 loss) | |
| I0525 03:26:41.758961 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00217896 (* 0.0909091 = 0.000198087 loss) | |
| I0525 03:26:41.758975 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00183909 (* 0.0909091 = 0.00016719 loss) | |
| I0525 03:26:41.758992 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00185611 (* 0.0909091 = 0.000168737 loss) | |
| I0525 03:26:41.759006 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00161079 (* 0.0909091 = 0.000146435 loss) | |
| I0525 03:26:41.759019 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00121633 (* 0.0909091 = 0.000110576 loss) | |
| I0525 03:26:41.759032 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 03:26:41.759042 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:26:41.759053 5272 solver.cpp:406] Test net output #149: total_confidence = 0.000325351 | |
| I0525 03:26:41.759064 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000452278 | |
| I0525 03:26:41.759088 5272 solver.cpp:338] Iteration 15000, Testing net (#1) | |
| I0525 03:27:39.668490 5272 solver.cpp:393] Test loss: 10.0552 | |
| I0525 03:27:39.668619 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0624187 | |
| I0525 03:27:39.668638 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.113 | |
| I0525 03:27:39.668653 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.11 | |
| I0525 03:27:39.668665 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.081 | |
| I0525 03:27:39.668678 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.173 | |
| I0525 03:27:39.668689 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.328 | |
| I0525 03:27:39.668702 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.447 | |
| I0525 03:27:39.668715 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.655 | |
| I0525 03:27:39.668726 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.825 | |
| I0525 03:27:39.668738 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.885 | |
| I0525 03:27:39.668751 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.902 | |
| I0525 03:27:39.668762 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.925 | |
| I0525 03:27:39.668774 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.942 | |
| I0525 03:27:39.668787 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.952 | |
| I0525 03:27:39.668798 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.961 | |
| I0525 03:27:39.668809 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.964 | |
| I0525 03:27:39.668822 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.982 | |
| I0525 03:27:39.668833 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.992 | |
| I0525 03:27:39.668844 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.993 | |
| I0525 03:27:39.668856 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.994 | |
| I0525 03:27:39.668869 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 0.998 | |
| I0525 03:27:39.668884 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 0.998 | |
| I0525 03:27:39.668896 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 0.998 | |
| I0525 03:27:39.668908 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.733365 | |
| I0525 03:27:39.668920 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.230859 | |
| I0525 03:27:39.668936 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.57206 (* 0.3 = 1.07162 loss) | |
| I0525 03:27:39.668951 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 1.04843 (* 0.3 = 0.31453 loss) | |
| I0525 03:27:39.668965 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 3.07091 (* 0.0272727 = 0.0837521 loss) | |
| I0525 03:27:39.668979 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.16638 (* 0.0272727 = 0.0863558 loss) | |
| I0525 03:27:39.668993 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.27769 (* 0.0272727 = 0.0893916 loss) | |
| I0525 03:27:39.669008 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.10727 (* 0.0272727 = 0.0847437 loss) | |
| I0525 03:27:39.669020 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 2.65066 (* 0.0272727 = 0.0722907 loss) | |
| I0525 03:27:39.669034 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.29253 (* 0.0272727 = 0.0625234 loss) | |
| I0525 03:27:39.669047 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.54685 (* 0.0272727 = 0.0421868 loss) | |
| I0525 03:27:39.669061 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.836177 (* 0.0272727 = 0.0228048 loss) | |
| I0525 03:27:39.669075 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.52312 (* 0.0272727 = 0.0142669 loss) | |
| I0525 03:27:39.669088 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.44185 (* 0.0272727 = 0.0120504 loss) | |
| I0525 03:27:39.669102 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.350504 (* 0.0272727 = 0.0095592 loss) | |
| I0525 03:27:39.669116 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.278084 (* 0.0272727 = 0.0075841 loss) | |
| I0525 03:27:39.669167 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.249441 (* 0.0272727 = 0.00680295 loss) | |
| I0525 03:27:39.669183 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.211546 (* 0.0272727 = 0.00576944 loss) | |
| I0525 03:27:39.669196 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.200438 (* 0.0272727 = 0.00546649 loss) | |
| I0525 03:27:39.669209 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.119281 (* 0.0272727 = 0.00325311 loss) | |
| I0525 03:27:39.669224 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.065333 (* 0.0272727 = 0.00178181 loss) | |
| I0525 03:27:39.669237 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0578593 (* 0.0272727 = 0.00157798 loss) | |
| I0525 03:27:39.669251 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0543099 (* 0.0272727 = 0.00148118 loss) | |
| I0525 03:27:39.669265 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.0208983 (* 0.0272727 = 0.000569953 loss) | |
| I0525 03:27:39.669280 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.0210826 (* 0.0272727 = 0.000574981 loss) | |
| I0525 03:27:39.669292 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.0193138 (* 0.0272727 = 0.000526741 loss) | |
| I0525 03:27:39.669304 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0620986 | |
| I0525 03:27:39.669317 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.103 | |
| I0525 03:27:39.669328 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.089 | |
| I0525 03:27:39.669340 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.09 | |
| I0525 03:27:39.669353 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.176 | |
| I0525 03:27:39.669363 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.339 | |
| I0525 03:27:39.669375 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.444 | |
| I0525 03:27:39.669386 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.656 | |
| I0525 03:27:39.669399 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.827 | |
| I0525 03:27:39.669409 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.885 | |
| I0525 03:27:39.669420 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.902 | |
| I0525 03:27:39.669432 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.925 | |
| I0525 03:27:39.669443 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.942 | |
| I0525 03:27:39.669456 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.952 | |
| I0525 03:27:39.669467 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.961 | |
| I0525 03:27:39.669479 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.964 | |
| I0525 03:27:39.669491 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.982 | |
| I0525 03:27:39.669502 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.992 | |
| I0525 03:27:39.669513 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.993 | |
| I0525 03:27:39.669524 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.994 | |
| I0525 03:27:39.669536 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 0.998 | |
| I0525 03:27:39.669548 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 0.998 | |
| I0525 03:27:39.669559 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 0.998 | |
| I0525 03:27:39.669570 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.733728 | |
| I0525 03:27:39.669581 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.23036 | |
| I0525 03:27:39.669595 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.57128 (* 0.3 = 1.07138 loss) | |
| I0525 03:27:39.669608 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.05228 (* 0.3 = 0.315684 loss) | |
| I0525 03:27:39.669622 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 3.03695 (* 0.0272727 = 0.0828259 loss) | |
| I0525 03:27:39.669639 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.1255 (* 0.0272727 = 0.085241 loss) | |
| I0525 03:27:39.669664 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.24062 (* 0.0272727 = 0.0883805 loss) | |
| I0525 03:27:39.669679 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 3.07091 (* 0.0272727 = 0.083752 loss) | |
| I0525 03:27:39.669693 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 2.60847 (* 0.0272727 = 0.0711402 loss) | |
| I0525 03:27:39.669706 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.26639 (* 0.0272727 = 0.0618107 loss) | |
| I0525 03:27:39.669720 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.52486 (* 0.0272727 = 0.041587 loss) | |
| I0525 03:27:39.669734 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.830165 (* 0.0272727 = 0.0226409 loss) | |
| I0525 03:27:39.669747 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.518391 (* 0.0272727 = 0.0141379 loss) | |
| I0525 03:27:39.669760 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.435276 (* 0.0272727 = 0.0118712 loss) | |
| I0525 03:27:39.669775 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.345118 (* 0.0272727 = 0.0094123 loss) | |
| I0525 03:27:39.669788 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.275516 (* 0.0272727 = 0.00751408 loss) | |
| I0525 03:27:39.669801 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.250039 (* 0.0272727 = 0.00681926 loss) | |
| I0525 03:27:39.669816 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.211915 (* 0.0272727 = 0.0057795 loss) | |
| I0525 03:27:39.669828 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.207491 (* 0.0272727 = 0.00565885 loss) | |
| I0525 03:27:39.669842 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.124061 (* 0.0272727 = 0.00338349 loss) | |
| I0525 03:27:39.669857 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0651923 (* 0.0272727 = 0.00177797 loss) | |
| I0525 03:27:39.669870 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0612407 (* 0.0272727 = 0.0016702 loss) | |
| I0525 03:27:39.669884 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0541983 (* 0.0272727 = 0.00147814 loss) | |
| I0525 03:27:39.669898 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0207097 (* 0.0272727 = 0.000564811 loss) | |
| I0525 03:27:39.669912 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.0202217 (* 0.0272727 = 0.0005515 loss) | |
| I0525 03:27:39.669929 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.0223832 (* 0.0272727 = 0.000610452 loss) | |
| I0525 03:27:39.669942 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0802786 | |
| I0525 03:27:39.669953 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.1 | |
| I0525 03:27:39.669965 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.097 | |
| I0525 03:27:39.669976 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.087 | |
| I0525 03:27:39.669988 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.175 | |
| I0525 03:27:39.669999 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.353 | |
| I0525 03:27:39.670011 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.447 | |
| I0525 03:27:39.670022 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.66 | |
| I0525 03:27:39.670033 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.826 | |
| I0525 03:27:39.670045 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.887 | |
| I0525 03:27:39.670056 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.904 | |
| I0525 03:27:39.670068 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.925 | |
| I0525 03:27:39.670079 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.942 | |
| I0525 03:27:39.670090 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.952 | |
| I0525 03:27:39.670101 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.961 | |
| I0525 03:27:39.670114 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.964 | |
| I0525 03:27:39.670121 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.982 | |
| I0525 03:27:39.670142 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.992 | |
| I0525 03:27:39.670156 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.993 | |
| I0525 03:27:39.670166 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.994 | |
| I0525 03:27:39.670178 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 0.998 | |
| I0525 03:27:39.670189 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 0.998 | |
| I0525 03:27:39.670200 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 0.998 | |
| I0525 03:27:39.670212 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.737228 | |
| I0525 03:27:39.670223 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.246684 | |
| I0525 03:27:39.670238 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.17301 (* 1 = 3.17301 loss) | |
| I0525 03:27:39.670251 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.9571 (* 1 = 0.9571 loss) | |
| I0525 03:27:39.670264 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 2.91385 (* 0.0909091 = 0.264895 loss) | |
| I0525 03:27:39.670279 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 2.99934 (* 0.0909091 = 0.272667 loss) | |
| I0525 03:27:39.670291 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.10188 (* 0.0909091 = 0.281989 loss) | |
| I0525 03:27:39.670305 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 2.97285 (* 0.0909091 = 0.270259 loss) | |
| I0525 03:27:39.670318 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.50756 (* 0.0909091 = 0.22796 loss) | |
| I0525 03:27:39.670331 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.14288 (* 0.0909091 = 0.194808 loss) | |
| I0525 03:27:39.670344 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.41917 (* 0.0909091 = 0.129016 loss) | |
| I0525 03:27:39.670358 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.779182 (* 0.0909091 = 0.0708348 loss) | |
| I0525 03:27:39.670372 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.478201 (* 0.0909091 = 0.0434728 loss) | |
| I0525 03:27:39.670385 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.402988 (* 0.0909091 = 0.0366353 loss) | |
| I0525 03:27:39.670398 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.312749 (* 0.0909091 = 0.0284317 loss) | |
| I0525 03:27:39.670413 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.245422 (* 0.0909091 = 0.0223111 loss) | |
| I0525 03:27:39.670425 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.22795 (* 0.0909091 = 0.0207227 loss) | |
| I0525 03:27:39.670439 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.191202 (* 0.0909091 = 0.017382 loss) | |
| I0525 03:27:39.670452 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.184095 (* 0.0909091 = 0.0167359 loss) | |
| I0525 03:27:39.670465 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.104678 (* 0.0909091 = 0.00951615 loss) | |
| I0525 03:27:39.670480 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.0529746 (* 0.0909091 = 0.00481588 loss) | |
| I0525 03:27:39.670492 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0536157 (* 0.0909091 = 0.00487415 loss) | |
| I0525 03:27:39.670506 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0490098 (* 0.0909091 = 0.00445544 loss) | |
| I0525 03:27:39.670519 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.0208882 (* 0.0909091 = 0.00189893 loss) | |
| I0525 03:27:39.670533 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0239523 (* 0.0909091 = 0.00217748 loss) | |
| I0525 03:27:39.670547 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.023586 (* 0.0909091 = 0.00214418 loss) | |
| I0525 03:27:39.670558 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 03:27:39.670569 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:27:39.670580 5272 solver.cpp:406] Test net output #149: total_confidence = 0.000267173 | |
| I0525 03:27:39.670600 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000348047 | |
| I0525 03:27:40.027756 5272 solver.cpp:229] Iteration 15000, loss = 9.99141 | |
| I0525 03:27:40.027820 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.06 | |
| I0525 03:27:40.027839 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 03:27:40.027853 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0525 03:27:40.027865 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0525 03:27:40.027878 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 03:27:40.027891 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 03:27:40.027904 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 03:27:40.027916 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 03:27:40.027930 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 03:27:40.027941 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:27:40.027953 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:27:40.027966 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:27:40.027978 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:27:40.027990 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:27:40.028002 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:27:40.028014 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:27:40.028025 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:27:40.028038 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:27:40.028049 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:27:40.028061 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:27:40.028074 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:27:40.028086 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:27:40.028098 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:27:40.028110 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0525 03:27:40.028121 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.26 | |
| I0525 03:27:40.028138 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.2736 (* 0.3 = 0.98208 loss) | |
| I0525 03:27:40.028152 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.989012 (* 0.3 = 0.296704 loss) | |
| I0525 03:27:40.028167 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.39907 (* 0.0272727 = 0.0927019 loss) | |
| I0525 03:27:40.028182 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.10783 (* 0.0272727 = 0.0847589 loss) | |
| I0525 03:27:40.028195 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.33076 (* 0.0272727 = 0.0908388 loss) | |
| I0525 03:27:40.028209 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.86844 (* 0.0272727 = 0.0782303 loss) | |
| I0525 03:27:40.028223 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.05794 (* 0.0272727 = 0.0833984 loss) | |
| I0525 03:27:40.028237 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.12607 (* 0.0272727 = 0.0852565 loss) | |
| I0525 03:27:40.028251 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.4996 (* 0.0272727 = 0.0408981 loss) | |
| I0525 03:27:40.028265 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.4426 (* 0.0272727 = 0.0393437 loss) | |
| I0525 03:27:40.028283 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.239513 (* 0.0272727 = 0.00653216 loss) | |
| I0525 03:27:40.028298 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.117337 (* 0.0272727 = 0.00320009 loss) | |
| I0525 03:27:40.028313 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.104551 (* 0.0272727 = 0.0028514 loss) | |
| I0525 03:27:40.028360 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.121336 (* 0.0272727 = 0.00330917 loss) | |
| I0525 03:27:40.028378 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0573485 (* 0.0272727 = 0.00156405 loss) | |
| I0525 03:27:40.028391 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0575126 (* 0.0272727 = 0.00156853 loss) | |
| I0525 03:27:40.028406 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0396513 (* 0.0272727 = 0.0010814 loss) | |
| I0525 03:27:40.028420 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0275282 (* 0.0272727 = 0.000750768 loss) | |
| I0525 03:27:40.028434 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0149263 (* 0.0272727 = 0.000407081 loss) | |
| I0525 03:27:40.028448 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0229183 (* 0.0272727 = 0.000625046 loss) | |
| I0525 03:27:40.028462 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0115301 (* 0.0272727 = 0.000314458 loss) | |
| I0525 03:27:40.028477 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00699912 (* 0.0272727 = 0.000190885 loss) | |
| I0525 03:27:40.028491 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00627207 (* 0.0272727 = 0.000171057 loss) | |
| I0525 03:27:40.028506 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00738677 (* 0.0272727 = 0.000201457 loss) | |
| I0525 03:27:40.028517 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.06 | |
| I0525 03:27:40.028530 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 03:27:40.028542 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 03:27:40.028554 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 03:27:40.028566 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 03:27:40.028578 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 03:27:40.028590 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.125 | |
| I0525 03:27:40.028602 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 03:27:40.028614 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 03:27:40.028626 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:27:40.028638 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:27:40.028650 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:27:40.028661 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:27:40.028673 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:27:40.028684 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:27:40.028697 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:27:40.028708 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:27:40.028719 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:27:40.028731 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:27:40.028743 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:27:40.028758 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:27:40.028769 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:27:40.028781 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:27:40.028792 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0525 03:27:40.028805 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.22 | |
| I0525 03:27:40.028820 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.29526 (* 0.3 = 0.988578 loss) | |
| I0525 03:27:40.028833 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.07101 (* 0.3 = 0.321303 loss) | |
| I0525 03:27:40.028847 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.61301 (* 0.0272727 = 0.0985367 loss) | |
| I0525 03:27:40.028873 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.70135 (* 0.0272727 = 0.100946 loss) | |
| I0525 03:27:40.028888 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 2.83592 (* 0.0272727 = 0.0773433 loss) | |
| I0525 03:27:40.028903 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 2.88441 (* 0.0272727 = 0.0786658 loss) | |
| I0525 03:27:40.028916 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.9348 (* 0.0272727 = 0.08004 loss) | |
| I0525 03:27:40.028930 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.67525 (* 0.0272727 = 0.100234 loss) | |
| I0525 03:27:40.028944 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.33968 (* 0.0272727 = 0.0365367 loss) | |
| I0525 03:27:40.028957 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.26109 (* 0.0272727 = 0.0343934 loss) | |
| I0525 03:27:40.028971 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.174744 (* 0.0272727 = 0.00476574 loss) | |
| I0525 03:27:40.028985 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0740329 (* 0.0272727 = 0.00201908 loss) | |
| I0525 03:27:40.029000 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0630543 (* 0.0272727 = 0.00171966 loss) | |
| I0525 03:27:40.029014 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0481258 (* 0.0272727 = 0.00131252 loss) | |
| I0525 03:27:40.029028 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0298549 (* 0.0272727 = 0.000814225 loss) | |
| I0525 03:27:40.029042 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0249457 (* 0.0272727 = 0.000680338 loss) | |
| I0525 03:27:40.029057 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0189401 (* 0.0272727 = 0.000516549 loss) | |
| I0525 03:27:40.029070 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0100057 (* 0.0272727 = 0.000272882 loss) | |
| I0525 03:27:40.029084 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00585015 (* 0.0272727 = 0.00015955 loss) | |
| I0525 03:27:40.029098 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00535054 (* 0.0272727 = 0.000145924 loss) | |
| I0525 03:27:40.029112 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0099426 (* 0.0272727 = 0.000271162 loss) | |
| I0525 03:27:40.029142 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00334323 (* 0.0272727 = 9.11789e-05 loss) | |
| I0525 03:27:40.029157 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00429528 (* 0.0272727 = 0.000117144 loss) | |
| I0525 03:27:40.029171 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00301502 (* 0.0272727 = 8.22278e-05 loss) | |
| I0525 03:27:40.029184 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.04 | |
| I0525 03:27:40.029196 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 03:27:40.029208 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 03:27:40.029220 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 03:27:40.029232 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 03:27:40.029244 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 03:27:40.029256 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0525 03:27:40.029268 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 03:27:40.029280 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 03:27:40.029292 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:27:40.029304 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:27:40.029316 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:27:40.029331 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:27:40.029343 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:27:40.029356 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:27:40.029378 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:27:40.029392 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:27:40.029404 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:27:40.029417 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:27:40.029428 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:27:40.029439 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:27:40.029451 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:27:40.029464 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:27:40.029474 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.727273 | |
| I0525 03:27:40.029484 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.22 | |
| I0525 03:27:40.029495 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.23706 (* 1 = 3.23706 loss) | |
| I0525 03:27:40.029508 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.972268 (* 1 = 0.972268 loss) | |
| I0525 03:27:40.029522 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.45488 (* 0.0909091 = 0.31408 loss) | |
| I0525 03:27:40.029536 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.35645 (* 0.0909091 = 0.305132 loss) | |
| I0525 03:27:40.029551 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.15619 (* 0.0909091 = 0.286926 loss) | |
| I0525 03:27:40.029564 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.8836 (* 0.0909091 = 0.262146 loss) | |
| I0525 03:27:40.029578 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.85568 (* 0.0909091 = 0.259607 loss) | |
| I0525 03:27:40.029592 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.18089 (* 0.0909091 = 0.289172 loss) | |
| I0525 03:27:40.029606 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.31532 (* 0.0909091 = 0.119575 loss) | |
| I0525 03:27:40.029620 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.07583 (* 0.0909091 = 0.0978026 loss) | |
| I0525 03:27:40.029634 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.244441 (* 0.0909091 = 0.0222219 loss) | |
| I0525 03:27:40.029649 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.116614 (* 0.0909091 = 0.0106013 loss) | |
| I0525 03:27:40.029662 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0785128 (* 0.0909091 = 0.00713753 loss) | |
| I0525 03:27:40.029676 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0559512 (* 0.0909091 = 0.00508647 loss) | |
| I0525 03:27:40.029691 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0417427 (* 0.0909091 = 0.00379479 loss) | |
| I0525 03:27:40.029706 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0337461 (* 0.0909091 = 0.00306783 loss) | |
| I0525 03:27:40.029719 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.03011 (* 0.0909091 = 0.00273728 loss) | |
| I0525 03:27:40.029733 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0166098 (* 0.0909091 = 0.00150999 loss) | |
| I0525 03:27:40.029747 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00840129 (* 0.0909091 = 0.000763754 loss) | |
| I0525 03:27:40.029762 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00434437 (* 0.0909091 = 0.000394943 loss) | |
| I0525 03:27:40.029775 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00304991 (* 0.0909091 = 0.000277265 loss) | |
| I0525 03:27:40.029789 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00424968 (* 0.0909091 = 0.000386334 loss) | |
| I0525 03:27:40.029806 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00223896 (* 0.0909091 = 0.000203542 loss) | |
| I0525 03:27:40.029820 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00168685 (* 0.0909091 = 0.00015335 loss) | |
| I0525 03:27:40.029834 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:27:40.029855 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:27:40.029867 5272 solver.cpp:245] Train net output #149: total_confidence = 7.97902e-06 | |
| I0525 03:27:40.029880 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.15929e-05 | |
| I0525 03:27:40.029892 5272 sgd_solver.cpp:106] Iteration 15000, lr = 0.001 | |
| I0525 03:27:47.314586 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8799 > 30) by scale factor 0.971506 | |
| I0525 03:30:12.770532 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.4854 > 30) by scale factor 0.822247 | |
| I0525 03:31:53.544445 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 49.1603 > 30) by scale factor 0.610249 | |
| I0525 03:33:45.122551 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.3631 > 30) by scale factor 0.98804 | |
| I0525 03:34:04.767937 5272 solver.cpp:229] Iteration 15500, loss = 9.9276 | |
| I0525 03:34:04.768004 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.142857 | |
| I0525 03:34:04.768023 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 03:34:04.768035 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 03:34:04.768049 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.375 | |
| I0525 03:34:04.768061 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 03:34:04.768074 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 03:34:04.768086 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 03:34:04.768100 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 03:34:04.768111 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 03:34:04.768124 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:34:04.768136 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:34:04.768149 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:34:04.768162 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:34:04.768173 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:34:04.768185 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:34:04.768196 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:34:04.768208 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:34:04.768220 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:34:04.768232 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:34:04.768244 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:34:04.768256 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:34:04.768268 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:34:04.768280 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:34:04.768292 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0525 03:34:04.768304 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.244898 | |
| I0525 03:34:04.768321 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.25834 (* 0.3 = 0.977501 loss) | |
| I0525 03:34:04.768335 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.10535 (* 0.3 = 0.331604 loss) | |
| I0525 03:34:04.768349 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.3001 (* 0.0272727 = 0.0900027 loss) | |
| I0525 03:34:04.768363 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.08349 (* 0.0272727 = 0.0840951 loss) | |
| I0525 03:34:04.768378 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 2.75727 (* 0.0272727 = 0.0751982 loss) | |
| I0525 03:34:04.768393 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.85069 (* 0.0272727 = 0.105019 loss) | |
| I0525 03:34:04.768406 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.90476 (* 0.0272727 = 0.0792207 loss) | |
| I0525 03:34:04.768420 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.76213 (* 0.0272727 = 0.0753307 loss) | |
| I0525 03:34:04.768434 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.36091 (* 0.0272727 = 0.0643885 loss) | |
| I0525 03:34:04.768447 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.724624 (* 0.0272727 = 0.0197625 loss) | |
| I0525 03:34:04.768462 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.144901 (* 0.0272727 = 0.00395185 loss) | |
| I0525 03:34:04.768476 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.141948 (* 0.0272727 = 0.0038713 loss) | |
| I0525 03:34:04.768491 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.137897 (* 0.0272727 = 0.00376082 loss) | |
| I0525 03:34:04.768506 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0885088 (* 0.0272727 = 0.00241388 loss) | |
| I0525 03:34:04.768553 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0505395 (* 0.0272727 = 0.00137835 loss) | |
| I0525 03:34:04.768568 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0356555 (* 0.0272727 = 0.000972422 loss) | |
| I0525 03:34:04.768584 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0222831 (* 0.0272727 = 0.000607721 loss) | |
| I0525 03:34:04.768597 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0364462 (* 0.0272727 = 0.000993988 loss) | |
| I0525 03:34:04.768615 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0284699 (* 0.0272727 = 0.000776451 loss) | |
| I0525 03:34:04.768630 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0139512 (* 0.0272727 = 0.000380486 loss) | |
| I0525 03:34:04.768646 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0154336 (* 0.0272727 = 0.000420917 loss) | |
| I0525 03:34:04.768659 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00994675 (* 0.0272727 = 0.000271275 loss) | |
| I0525 03:34:04.768673 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0098096 (* 0.0272727 = 0.000267534 loss) | |
| I0525 03:34:04.768687 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00799504 (* 0.0272727 = 0.000218047 loss) | |
| I0525 03:34:04.768699 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0408163 | |
| I0525 03:34:04.768712 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 03:34:04.768724 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 03:34:04.768735 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 03:34:04.768749 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 03:34:04.768761 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 03:34:04.768774 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 03:34:04.768785 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 03:34:04.768797 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 03:34:04.768810 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:34:04.768821 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:34:04.768832 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:34:04.768844 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:34:04.768856 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:34:04.768867 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:34:04.768878 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:34:04.768890 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:34:04.768901 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:34:04.768913 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:34:04.768924 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:34:04.768935 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:34:04.768947 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:34:04.768959 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:34:04.768970 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.721591 | |
| I0525 03:34:04.768981 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.183673 | |
| I0525 03:34:04.768996 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.29373 (* 0.3 = 0.988119 loss) | |
| I0525 03:34:04.769009 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.03251 (* 0.3 = 0.309752 loss) | |
| I0525 03:34:04.769022 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.03227 (* 0.0272727 = 0.0826982 loss) | |
| I0525 03:34:04.769037 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.15294 (* 0.0272727 = 0.0859893 loss) | |
| I0525 03:34:04.769062 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.02747 (* 0.0272727 = 0.0825673 loss) | |
| I0525 03:34:04.769075 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.59206 (* 0.0272727 = 0.0979653 loss) | |
| I0525 03:34:04.769089 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.52609 (* 0.0272727 = 0.0688933 loss) | |
| I0525 03:34:04.769104 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.33103 (* 0.0272727 = 0.0635735 loss) | |
| I0525 03:34:04.769134 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.3001 (* 0.0272727 = 0.06273 loss) | |
| I0525 03:34:04.769151 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.656524 (* 0.0272727 = 0.0179052 loss) | |
| I0525 03:34:04.769166 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0744496 (* 0.0272727 = 0.00203044 loss) | |
| I0525 03:34:04.769181 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0658119 (* 0.0272727 = 0.00179487 loss) | |
| I0525 03:34:04.769194 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0394265 (* 0.0272727 = 0.00107527 loss) | |
| I0525 03:34:04.769209 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0314298 (* 0.0272727 = 0.000857176 loss) | |
| I0525 03:34:04.769223 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0101711 (* 0.0272727 = 0.000277392 loss) | |
| I0525 03:34:04.769237 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0157271 (* 0.0272727 = 0.00042892 loss) | |
| I0525 03:34:04.769251 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0230788 (* 0.0272727 = 0.000629423 loss) | |
| I0525 03:34:04.769265 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00992693 (* 0.0272727 = 0.000270734 loss) | |
| I0525 03:34:04.769279 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0188802 (* 0.0272727 = 0.000514913 loss) | |
| I0525 03:34:04.769294 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0167816 (* 0.0272727 = 0.00045768 loss) | |
| I0525 03:34:04.769307 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0138034 (* 0.0272727 = 0.000376456 loss) | |
| I0525 03:34:04.769321 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0112452 (* 0.0272727 = 0.000306688 loss) | |
| I0525 03:34:04.769335 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00701348 (* 0.0272727 = 0.000191277 loss) | |
| I0525 03:34:04.769348 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00649548 (* 0.0272727 = 0.00017715 loss) | |
| I0525 03:34:04.769361 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0816327 | |
| I0525 03:34:04.769373 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 03:34:04.769384 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 03:34:04.769397 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 03:34:04.769408 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 03:34:04.769419 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 03:34:04.769431 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 03:34:04.769443 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 03:34:04.769455 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 03:34:04.769467 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:34:04.769479 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:34:04.769490 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:34:04.769501 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:34:04.769512 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:34:04.769525 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:34:04.769536 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:34:04.769558 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:34:04.769572 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:34:04.769583 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:34:04.769595 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:34:04.769608 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:34:04.769618 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:34:04.769630 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:34:04.769641 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0525 03:34:04.769654 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.285714 | |
| I0525 03:34:04.769670 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.80814 (* 1 = 2.80814 loss) | |
| I0525 03:34:04.769685 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.877005 (* 1 = 0.877005 loss) | |
| I0525 03:34:04.769698 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.61296 (* 0.0909091 = 0.237542 loss) | |
| I0525 03:34:04.769712 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.8794 (* 0.0909091 = 0.261763 loss) | |
| I0525 03:34:04.769726 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.77366 (* 0.0909091 = 0.252151 loss) | |
| I0525 03:34:04.769740 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.54035 (* 0.0909091 = 0.32185 loss) | |
| I0525 03:34:04.769753 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.38097 (* 0.0909091 = 0.216452 loss) | |
| I0525 03:34:04.769767 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.08633 (* 0.0909091 = 0.189667 loss) | |
| I0525 03:34:04.769781 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.06382 (* 0.0909091 = 0.18762 loss) | |
| I0525 03:34:04.769798 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.536527 (* 0.0909091 = 0.0487752 loss) | |
| I0525 03:34:04.769811 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.113475 (* 0.0909091 = 0.0103159 loss) | |
| I0525 03:34:04.769825 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0794687 (* 0.0909091 = 0.00722443 loss) | |
| I0525 03:34:04.769840 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0626262 (* 0.0909091 = 0.00569329 loss) | |
| I0525 03:34:04.769855 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0369124 (* 0.0909091 = 0.00335567 loss) | |
| I0525 03:34:04.769868 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0215954 (* 0.0909091 = 0.00196322 loss) | |
| I0525 03:34:04.769881 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0150633 (* 0.0909091 = 0.00136939 loss) | |
| I0525 03:34:04.769896 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0134135 (* 0.0909091 = 0.00121941 loss) | |
| I0525 03:34:04.769909 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00711239 (* 0.0909091 = 0.000646581 loss) | |
| I0525 03:34:04.769927 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00650103 (* 0.0909091 = 0.000591003 loss) | |
| I0525 03:34:04.769940 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00591476 (* 0.0909091 = 0.000537705 loss) | |
| I0525 03:34:04.769954 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00230234 (* 0.0909091 = 0.000209304 loss) | |
| I0525 03:34:04.769968 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00236752 (* 0.0909091 = 0.000215229 loss) | |
| I0525 03:34:04.769982 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00231042 (* 0.0909091 = 0.000210038 loss) | |
| I0525 03:34:04.769995 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00092201 (* 0.0909091 = 8.38191e-05 loss) | |
| I0525 03:34:04.770009 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:34:04.770020 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:34:04.770041 5272 solver.cpp:245] Train net output #149: total_confidence = 4.96809e-05 | |
| I0525 03:34:04.770054 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000163431 | |
| I0525 03:34:04.770067 5272 sgd_solver.cpp:106] Iteration 15500, lr = 0.001 | |
| I0525 03:35:36.707347 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 52.9072 > 30) by scale factor 0.567031 | |
| I0525 03:39:35.287448 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.995 > 30) by scale factor 0.937647 | |
| I0525 03:40:29.610981 5272 solver.cpp:229] Iteration 16000, loss = 10.0087 | |
| I0525 03:40:29.611110 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0847458 | |
| I0525 03:40:29.611132 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 03:40:29.611145 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 03:40:29.611158 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0525 03:40:29.611171 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 03:40:29.611183 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 03:40:29.611196 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 03:40:29.611209 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.375 | |
| I0525 03:40:29.611222 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.375 | |
| I0525 03:40:29.611234 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.625 | |
| I0525 03:40:29.611248 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 03:40:29.611261 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 03:40:29.611274 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 03:40:29.611286 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 03:40:29.611299 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:40:29.611310 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:40:29.611321 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:40:29.611333 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:40:29.611345 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:40:29.611356 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:40:29.611368 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:40:29.611380 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:40:29.611392 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:40:29.611404 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.6875 | |
| I0525 03:40:29.611416 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.186441 | |
| I0525 03:40:29.611433 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.13995 (* 0.3 = 0.941986 loss) | |
| I0525 03:40:29.611448 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.24203 (* 0.3 = 0.372609 loss) | |
| I0525 03:40:29.611462 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.38016 (* 0.0272727 = 0.0921863 loss) | |
| I0525 03:40:29.611476 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 2.99163 (* 0.0272727 = 0.0815898 loss) | |
| I0525 03:40:29.611490 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 2.99546 (* 0.0272727 = 0.0816943 loss) | |
| I0525 03:40:29.611505 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.01218 (* 0.0272727 = 0.0821502 loss) | |
| I0525 03:40:29.611517 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.17431 (* 0.0272727 = 0.0592995 loss) | |
| I0525 03:40:29.611532 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.30817 (* 0.0272727 = 0.06295 loss) | |
| I0525 03:40:29.611546 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.56759 (* 0.0272727 = 0.0700251 loss) | |
| I0525 03:40:29.611560 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 2.54102 (* 0.0272727 = 0.0693007 loss) | |
| I0525 03:40:29.611573 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.50636 (* 0.0272727 = 0.0410824 loss) | |
| I0525 03:40:29.611588 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.648441 (* 0.0272727 = 0.0176847 loss) | |
| I0525 03:40:29.611603 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.673985 (* 0.0272727 = 0.0183814 loss) | |
| I0525 03:40:29.611615 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.659943 (* 0.0272727 = 0.0179984 loss) | |
| I0525 03:40:29.611629 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.406591 (* 0.0272727 = 0.0110888 loss) | |
| I0525 03:40:29.611665 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.134549 (* 0.0272727 = 0.00366951 loss) | |
| I0525 03:40:29.611680 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0984001 (* 0.0272727 = 0.00268364 loss) | |
| I0525 03:40:29.611695 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0292249 (* 0.0272727 = 0.000797044 loss) | |
| I0525 03:40:29.611708 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0133283 (* 0.0272727 = 0.000363499 loss) | |
| I0525 03:40:29.611722 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0276383 (* 0.0272727 = 0.000753773 loss) | |
| I0525 03:40:29.611737 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0117294 (* 0.0272727 = 0.000319894 loss) | |
| I0525 03:40:29.611750 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0065907 (* 0.0272727 = 0.000179746 loss) | |
| I0525 03:40:29.611764 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00492645 (* 0.0272727 = 0.000134358 loss) | |
| I0525 03:40:29.611778 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00552311 (* 0.0272727 = 0.00015063 loss) | |
| I0525 03:40:29.611791 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.118644 | |
| I0525 03:40:29.611804 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 03:40:29.611816 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0525 03:40:29.611829 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 03:40:29.611840 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 03:40:29.611852 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 03:40:29.611865 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 03:40:29.611879 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.375 | |
| I0525 03:40:29.611892 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.375 | |
| I0525 03:40:29.611904 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.625 | |
| I0525 03:40:29.611917 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 03:40:29.611928 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 03:40:29.611940 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 03:40:29.611953 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 03:40:29.611964 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:40:29.611976 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:40:29.611987 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:40:29.611999 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:40:29.612011 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:40:29.612022 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:40:29.612035 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:40:29.612046 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:40:29.612057 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:40:29.612068 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.704545 | |
| I0525 03:40:29.612081 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.237288 | |
| I0525 03:40:29.612094 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.11685 (* 0.3 = 0.935056 loss) | |
| I0525 03:40:29.612108 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.14148 (* 0.3 = 0.342443 loss) | |
| I0525 03:40:29.612125 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.18168 (* 0.0272727 = 0.0867731 loss) | |
| I0525 03:40:29.612140 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.27047 (* 0.0272727 = 0.0891946 loss) | |
| I0525 03:40:29.612165 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 2.80781 (* 0.0272727 = 0.0765766 loss) | |
| I0525 03:40:29.612180 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.03835 (* 0.0272727 = 0.0828641 loss) | |
| I0525 03:40:29.612195 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.28662 (* 0.0272727 = 0.0623623 loss) | |
| I0525 03:40:29.612208 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.45987 (* 0.0272727 = 0.0670874 loss) | |
| I0525 03:40:29.612221 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.68855 (* 0.0272727 = 0.0733241 loss) | |
| I0525 03:40:29.612236 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 2.45544 (* 0.0272727 = 0.0669664 loss) | |
| I0525 03:40:29.612249 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.69385 (* 0.0272727 = 0.046196 loss) | |
| I0525 03:40:29.612263 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.686413 (* 0.0272727 = 0.0187203 loss) | |
| I0525 03:40:29.612277 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.51738 (* 0.0272727 = 0.0141104 loss) | |
| I0525 03:40:29.612290 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.664432 (* 0.0272727 = 0.0181209 loss) | |
| I0525 03:40:29.612304 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.421833 (* 0.0272727 = 0.0115045 loss) | |
| I0525 03:40:29.612318 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0986402 (* 0.0272727 = 0.00269019 loss) | |
| I0525 03:40:29.612332 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0634175 (* 0.0272727 = 0.00172957 loss) | |
| I0525 03:40:29.612346 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0645511 (* 0.0272727 = 0.00176048 loss) | |
| I0525 03:40:29.612361 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0275928 (* 0.0272727 = 0.000752531 loss) | |
| I0525 03:40:29.612375 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00864016 (* 0.0272727 = 0.000235641 loss) | |
| I0525 03:40:29.612390 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0239436 (* 0.0272727 = 0.000653007 loss) | |
| I0525 03:40:29.612403 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00545371 (* 0.0272727 = 0.000148738 loss) | |
| I0525 03:40:29.612417 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00851153 (* 0.0272727 = 0.000232133 loss) | |
| I0525 03:40:29.612432 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0135096 (* 0.0272727 = 0.000368445 loss) | |
| I0525 03:40:29.612444 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.169492 | |
| I0525 03:40:29.612457 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 03:40:29.612468 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 03:40:29.612480 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.25 | |
| I0525 03:40:29.612494 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 03:40:29.612506 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 03:40:29.612519 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 03:40:29.612530 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0525 03:40:29.612542 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.375 | |
| I0525 03:40:29.612553 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0525 03:40:29.612565 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 03:40:29.612577 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 03:40:29.612589 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 03:40:29.612601 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 03:40:29.612612 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:40:29.612624 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:40:29.612637 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:40:29.612658 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:40:29.612670 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:40:29.612681 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:40:29.612694 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:40:29.612705 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:40:29.612716 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:40:29.612727 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.721591 | |
| I0525 03:40:29.612740 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.271186 | |
| I0525 03:40:29.612753 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.91297 (* 1 = 2.91297 loss) | |
| I0525 03:40:29.612768 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.08626 (* 1 = 1.08626 loss) | |
| I0525 03:40:29.612782 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.87178 (* 0.0909091 = 0.261071 loss) | |
| I0525 03:40:29.612795 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.88061 (* 0.0909091 = 0.261874 loss) | |
| I0525 03:40:29.612809 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.44347 (* 0.0909091 = 0.222133 loss) | |
| I0525 03:40:29.612823 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.69387 (* 0.0909091 = 0.244898 loss) | |
| I0525 03:40:29.612838 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 1.89877 (* 0.0909091 = 0.172615 loss) | |
| I0525 03:40:29.612851 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.83026 (* 0.0909091 = 0.166387 loss) | |
| I0525 03:40:29.612864 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.52894 (* 0.0909091 = 0.229904 loss) | |
| I0525 03:40:29.612879 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 2.30378 (* 0.0909091 = 0.209434 loss) | |
| I0525 03:40:29.612891 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 1.31315 (* 0.0909091 = 0.119377 loss) | |
| I0525 03:40:29.612905 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.59627 (* 0.0909091 = 0.0542064 loss) | |
| I0525 03:40:29.612920 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.610379 (* 0.0909091 = 0.055489 loss) | |
| I0525 03:40:29.612936 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.369548 (* 0.0909091 = 0.0335952 loss) | |
| I0525 03:40:29.612951 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.396128 (* 0.0909091 = 0.0360116 loss) | |
| I0525 03:40:29.612965 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0364827 (* 0.0909091 = 0.00331661 loss) | |
| I0525 03:40:29.612979 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0172525 (* 0.0909091 = 0.00156841 loss) | |
| I0525 03:40:29.612993 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0115012 (* 0.0909091 = 0.00104556 loss) | |
| I0525 03:40:29.613008 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0149426 (* 0.0909091 = 0.00135842 loss) | |
| I0525 03:40:29.613021 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00593184 (* 0.0909091 = 0.000539259 loss) | |
| I0525 03:40:29.613035 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00313401 (* 0.0909091 = 0.00028491 loss) | |
| I0525 03:40:29.613049 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0012625 (* 0.0909091 = 0.000114773 loss) | |
| I0525 03:40:29.613064 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000638181 (* 0.0909091 = 5.80165e-05 loss) | |
| I0525 03:40:29.613077 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000431056 (* 0.0909091 = 3.91869e-05 loss) | |
| I0525 03:40:29.613090 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:40:29.613101 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:40:29.613113 5272 solver.cpp:245] Train net output #149: total_confidence = 3.4372e-05 | |
| I0525 03:40:29.613150 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000203559 | |
| I0525 03:40:29.613167 5272 sgd_solver.cpp:106] Iteration 16000, lr = 0.001 | |
| I0525 03:41:05.377734 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.1014 > 30) by scale factor 0.934538 | |
| I0525 03:42:25.434659 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.3097 > 30) by scale factor 0.900638 | |
| I0525 03:45:13.289178 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.2507 > 30) by scale factor 0.959977 | |
| I0525 03:45:36.383708 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.2675 > 30) by scale factor 0.875466 | |
| I0525 03:45:41.779907 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.026 > 30) by scale factor 0.749513 | |
| I0525 03:46:54.617725 5272 solver.cpp:229] Iteration 16500, loss = 9.95266 | |
| I0525 03:46:54.617871 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.137255 | |
| I0525 03:46:54.617892 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 03:46:54.617907 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 03:46:54.617919 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 03:46:54.617933 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 03:46:54.617944 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 03:46:54.617957 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 03:46:54.617969 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 03:46:54.617981 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 03:46:54.617995 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 03:46:54.618006 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:46:54.618019 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:46:54.618032 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:46:54.618046 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:46:54.618057 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:46:54.618068 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:46:54.618080 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:46:54.618091 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:46:54.618104 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:46:54.618115 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:46:54.618127 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:46:54.618139 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:46:54.618151 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:46:54.618162 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0525 03:46:54.618175 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.352941 | |
| I0525 03:46:54.618191 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.05289 (* 0.3 = 0.915868 loss) | |
| I0525 03:46:54.618206 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.951113 (* 0.3 = 0.285334 loss) | |
| I0525 03:46:54.618219 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.58906 (* 0.0272727 = 0.0706107 loss) | |
| I0525 03:46:54.618233 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.02925 (* 0.0272727 = 0.082616 loss) | |
| I0525 03:46:54.618247 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.09303 (* 0.0272727 = 0.0843554 loss) | |
| I0525 03:46:54.618262 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.87012 (* 0.0272727 = 0.078276 loss) | |
| I0525 03:46:54.618276 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.73307 (* 0.0272727 = 0.0745383 loss) | |
| I0525 03:46:54.618289 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.92504 (* 0.0272727 = 0.0797738 loss) | |
| I0525 03:46:54.618304 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.14681 (* 0.0272727 = 0.0585494 loss) | |
| I0525 03:46:54.618317 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.862504 (* 0.0272727 = 0.0235228 loss) | |
| I0525 03:46:54.618332 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.46301 (* 0.0272727 = 0.0126276 loss) | |
| I0525 03:46:54.618346 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0802536 (* 0.0272727 = 0.00218874 loss) | |
| I0525 03:46:54.618361 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0752945 (* 0.0272727 = 0.00205349 loss) | |
| I0525 03:46:54.618376 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0823297 (* 0.0272727 = 0.00224535 loss) | |
| I0525 03:46:54.618389 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0323772 (* 0.0272727 = 0.000883014 loss) | |
| I0525 03:46:54.618425 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.025092 (* 0.0272727 = 0.000684328 loss) | |
| I0525 03:46:54.618441 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0317565 (* 0.0272727 = 0.000866085 loss) | |
| I0525 03:46:54.618455 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00680745 (* 0.0272727 = 0.000185658 loss) | |
| I0525 03:46:54.618468 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00657653 (* 0.0272727 = 0.00017936 loss) | |
| I0525 03:46:54.618482 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00475138 (* 0.0272727 = 0.000129583 loss) | |
| I0525 03:46:54.618496 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00481302 (* 0.0272727 = 0.000131264 loss) | |
| I0525 03:46:54.618511 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00348588 (* 0.0272727 = 9.50693e-05 loss) | |
| I0525 03:46:54.618525 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00229253 (* 0.0272727 = 6.25236e-05 loss) | |
| I0525 03:46:54.618538 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00296761 (* 0.0272727 = 8.09349e-05 loss) | |
| I0525 03:46:54.618551 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.137255 | |
| I0525 03:46:54.618563 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 03:46:54.618576 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 03:46:54.618587 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 03:46:54.618599 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 03:46:54.618612 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 03:46:54.618623 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 03:46:54.618635 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 03:46:54.618648 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 03:46:54.618659 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 03:46:54.618671 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:46:54.618683 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:46:54.618695 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:46:54.618706 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:46:54.618718 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:46:54.618731 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:46:54.618741 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:46:54.618753 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:46:54.618765 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:46:54.618777 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:46:54.618788 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:46:54.618800 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:46:54.618811 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:46:54.618823 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0525 03:46:54.618835 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.254902 | |
| I0525 03:46:54.618849 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.95707 (* 0.3 = 0.887122 loss) | |
| I0525 03:46:54.618863 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.945267 (* 0.3 = 0.28358 loss) | |
| I0525 03:46:54.618881 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.54671 (* 0.0272727 = 0.0694558 loss) | |
| I0525 03:46:54.618892 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.89973 (* 0.0272727 = 0.0790834 loss) | |
| I0525 03:46:54.618918 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.00472 (* 0.0272727 = 0.0819469 loss) | |
| I0525 03:46:54.618933 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 2.75665 (* 0.0272727 = 0.0751814 loss) | |
| I0525 03:46:54.618947 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.81052 (* 0.0272727 = 0.0766505 loss) | |
| I0525 03:46:54.618962 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.8664 (* 0.0272727 = 0.0781745 loss) | |
| I0525 03:46:54.618974 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.2521 (* 0.0272727 = 0.0614209 loss) | |
| I0525 03:46:54.618988 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.08122 (* 0.0272727 = 0.0294879 loss) | |
| I0525 03:46:54.619002 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.569229 (* 0.0272727 = 0.0155244 loss) | |
| I0525 03:46:54.619017 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0327131 (* 0.0272727 = 0.000892174 loss) | |
| I0525 03:46:54.619031 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00847852 (* 0.0272727 = 0.000231232 loss) | |
| I0525 03:46:54.619045 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00721386 (* 0.0272727 = 0.000196742 loss) | |
| I0525 03:46:54.619060 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0075548 (* 0.0272727 = 0.00020604 loss) | |
| I0525 03:46:54.619072 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00696494 (* 0.0272727 = 0.000189953 loss) | |
| I0525 03:46:54.619086 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00632296 (* 0.0272727 = 0.000172444 loss) | |
| I0525 03:46:54.619101 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0030857 (* 0.0272727 = 8.41554e-05 loss) | |
| I0525 03:46:54.619114 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00404422 (* 0.0272727 = 0.000110297 loss) | |
| I0525 03:46:54.619128 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00145885 (* 0.0272727 = 3.97869e-05 loss) | |
| I0525 03:46:54.619143 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00162843 (* 0.0272727 = 4.44118e-05 loss) | |
| I0525 03:46:54.619156 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.000827683 (* 0.0272727 = 2.25732e-05 loss) | |
| I0525 03:46:54.619170 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0014765 (* 0.0272727 = 4.02681e-05 loss) | |
| I0525 03:46:54.619184 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.000740876 (* 0.0272727 = 2.02057e-05 loss) | |
| I0525 03:46:54.619196 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.156863 | |
| I0525 03:46:54.619209 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0525 03:46:54.619220 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 03:46:54.619232 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 03:46:54.619245 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0525 03:46:54.619256 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 03:46:54.619268 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 03:46:54.619280 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 03:46:54.619292 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 03:46:54.619303 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 03:46:54.619315 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:46:54.619328 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:46:54.619338 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:46:54.619349 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:46:54.619361 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:46:54.619372 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:46:54.619385 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:46:54.619405 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:46:54.619417 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:46:54.619429 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:46:54.619441 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:46:54.619452 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:46:54.619463 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:46:54.619475 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.755682 | |
| I0525 03:46:54.619488 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.294118 | |
| I0525 03:46:54.619501 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.84986 (* 1 = 2.84986 loss) | |
| I0525 03:46:54.619515 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.876946 (* 1 = 0.876946 loss) | |
| I0525 03:46:54.619529 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.45544 (* 0.0909091 = 0.223222 loss) | |
| I0525 03:46:54.619544 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.90914 (* 0.0909091 = 0.264468 loss) | |
| I0525 03:46:54.619557 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.74934 (* 0.0909091 = 0.24994 loss) | |
| I0525 03:46:54.619571 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.55501 (* 0.0909091 = 0.232273 loss) | |
| I0525 03:46:54.619585 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.56215 (* 0.0909091 = 0.232923 loss) | |
| I0525 03:46:54.619598 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.70508 (* 0.0909091 = 0.245917 loss) | |
| I0525 03:46:54.619611 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.99877 (* 0.0909091 = 0.181706 loss) | |
| I0525 03:46:54.619626 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.856668 (* 0.0909091 = 0.0778789 loss) | |
| I0525 03:46:54.619639 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.471886 (* 0.0909091 = 0.0428987 loss) | |
| I0525 03:46:54.619649 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0738205 (* 0.0909091 = 0.00671095 loss) | |
| I0525 03:46:54.619659 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0239838 (* 0.0909091 = 0.00218034 loss) | |
| I0525 03:46:54.619674 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0142832 (* 0.0909091 = 0.00129847 loss) | |
| I0525 03:46:54.619688 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0163333 (* 0.0909091 = 0.00148484 loss) | |
| I0525 03:46:54.619702 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00940803 (* 0.0909091 = 0.000855275 loss) | |
| I0525 03:46:54.619716 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0101508 (* 0.0909091 = 0.000922797 loss) | |
| I0525 03:46:54.619729 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0125232 (* 0.0909091 = 0.00113848 loss) | |
| I0525 03:46:54.619743 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0125034 (* 0.0909091 = 0.00113667 loss) | |
| I0525 03:46:54.619757 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00544933 (* 0.0909091 = 0.000495393 loss) | |
| I0525 03:46:54.619771 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00342997 (* 0.0909091 = 0.000311816 loss) | |
| I0525 03:46:54.619784 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.0025867 (* 0.0909091 = 0.000235154 loss) | |
| I0525 03:46:54.619798 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000747743 (* 0.0909091 = 6.79766e-05 loss) | |
| I0525 03:46:54.619812 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000796393 (* 0.0909091 = 7.23994e-05 loss) | |
| I0525 03:46:54.619824 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:46:54.619835 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:46:54.619846 5272 solver.cpp:245] Train net output #149: total_confidence = 2.49174e-05 | |
| I0525 03:46:54.619868 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000139793 | |
| I0525 03:46:54.619881 5272 sgd_solver.cpp:106] Iteration 16500, lr = 0.001 | |
| I0525 03:48:02.751521 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.9711 > 30) by scale factor 0.883105 | |
| I0525 03:51:33.928478 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.0884 > 30) by scale factor 0.880064 | |
| I0525 03:51:46.240334 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.857 > 30) by scale factor 0.913046 | |
| I0525 03:52:44.756243 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.7742 > 30) by scale factor 0.974842 | |
| I0525 03:53:15.541491 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 34.6426 > 30) by scale factor 0.865986 | |
| I0525 03:53:19.792511 5272 solver.cpp:229] Iteration 17000, loss = 9.96095 | |
| I0525 03:53:19.792582 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0697674 | |
| I0525 03:53:19.792599 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 03:53:19.792613 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 03:53:19.792626 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 03:53:19.792639 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 03:53:19.792651 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 03:53:19.792665 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 03:53:19.792677 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 03:53:19.792692 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 03:53:19.792706 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:53:19.792717 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:53:19.792731 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:53:19.792742 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:53:19.792754 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:53:19.792767 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:53:19.792778 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:53:19.792790 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:53:19.792803 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:53:19.792814 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:53:19.792826 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:53:19.792839 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:53:19.792850 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:53:19.792861 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:53:19.792873 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0525 03:53:19.792886 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.255814 | |
| I0525 03:53:19.792901 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.98732 (* 0.3 = 0.896196 loss) | |
| I0525 03:53:19.792917 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.804036 (* 0.3 = 0.241211 loss) | |
| I0525 03:53:19.792930 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.80692 (* 0.0272727 = 0.0765524 loss) | |
| I0525 03:53:19.792944 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.44307 (* 0.0272727 = 0.0939018 loss) | |
| I0525 03:53:19.792958 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.4879 (* 0.0272727 = 0.0951245 loss) | |
| I0525 03:53:19.792973 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.81378 (* 0.0272727 = 0.0767393 loss) | |
| I0525 03:53:19.792986 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.50198 (* 0.0272727 = 0.0682359 loss) | |
| I0525 03:53:19.793000 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.06775 (* 0.0272727 = 0.0563931 loss) | |
| I0525 03:53:19.793015 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.24982 (* 0.0272727 = 0.0340859 loss) | |
| I0525 03:53:19.793028 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.702248 (* 0.0272727 = 0.0191522 loss) | |
| I0525 03:53:19.793045 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.00603944 (* 0.0272727 = 0.000164712 loss) | |
| I0525 03:53:19.793059 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00419402 (* 0.0272727 = 0.000114382 loss) | |
| I0525 03:53:19.793073 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.00299061 (* 0.0272727 = 8.1562e-05 loss) | |
| I0525 03:53:19.793088 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00218774 (* 0.0272727 = 5.96655e-05 loss) | |
| I0525 03:53:19.793156 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00257254 (* 0.0272727 = 7.01603e-05 loss) | |
| I0525 03:53:19.793172 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0014997 (* 0.0272727 = 4.0901e-05 loss) | |
| I0525 03:53:19.793187 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0029842 (* 0.0272727 = 8.13874e-05 loss) | |
| I0525 03:53:19.793201 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00378948 (* 0.0272727 = 0.000103349 loss) | |
| I0525 03:53:19.793215 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0018549 (* 0.0272727 = 5.05881e-05 loss) | |
| I0525 03:53:19.793229 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00257865 (* 0.0272727 = 7.03268e-05 loss) | |
| I0525 03:53:19.793243 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00162324 (* 0.0272727 = 4.42702e-05 loss) | |
| I0525 03:53:19.793257 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00189218 (* 0.0272727 = 5.1605e-05 loss) | |
| I0525 03:53:19.793272 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00166128 (* 0.0272727 = 4.53076e-05 loss) | |
| I0525 03:53:19.793287 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00406098 (* 0.0272727 = 0.000110754 loss) | |
| I0525 03:53:19.793298 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0697674 | |
| I0525 03:53:19.793311 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 03:53:19.793323 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 03:53:19.793334 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 03:53:19.793346 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.125 | |
| I0525 03:53:19.793359 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 03:53:19.793370 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 03:53:19.793382 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 03:53:19.793395 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 03:53:19.793406 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:53:19.793417 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:53:19.793428 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:53:19.793440 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:53:19.793452 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:53:19.793463 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:53:19.793474 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:53:19.793486 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:53:19.793498 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:53:19.793510 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:53:19.793521 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:53:19.793534 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:53:19.793545 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:53:19.793556 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:53:19.793567 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.772727 | |
| I0525 03:53:19.793581 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.348837 | |
| I0525 03:53:19.793593 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 2.82465 (* 0.3 = 0.847394 loss) | |
| I0525 03:53:19.793607 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.775539 (* 0.3 = 0.232662 loss) | |
| I0525 03:53:19.793622 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.63012 (* 0.0272727 = 0.0717305 loss) | |
| I0525 03:53:19.793635 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.39972 (* 0.0272727 = 0.0927197 loss) | |
| I0525 03:53:19.793661 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 2.95927 (* 0.0272727 = 0.0807074 loss) | |
| I0525 03:53:19.793676 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 2.6357 (* 0.0272727 = 0.0718829 loss) | |
| I0525 03:53:19.793690 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.51045 (* 0.0272727 = 0.0684667 loss) | |
| I0525 03:53:19.793704 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 1.90236 (* 0.0272727 = 0.0518826 loss) | |
| I0525 03:53:19.793717 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.15075 (* 0.0272727 = 0.0313841 loss) | |
| I0525 03:53:19.793735 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.732372 (* 0.0272727 = 0.0199738 loss) | |
| I0525 03:53:19.793750 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00747531 (* 0.0272727 = 0.000203872 loss) | |
| I0525 03:53:19.793763 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00712594 (* 0.0272727 = 0.000194344 loss) | |
| I0525 03:53:19.793776 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00469047 (* 0.0272727 = 0.000127922 loss) | |
| I0525 03:53:19.793792 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00326161 (* 0.0272727 = 8.89529e-05 loss) | |
| I0525 03:53:19.793805 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00365549 (* 0.0272727 = 9.96953e-05 loss) | |
| I0525 03:53:19.793818 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00643849 (* 0.0272727 = 0.000175595 loss) | |
| I0525 03:53:19.793833 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00476655 (* 0.0272727 = 0.000129997 loss) | |
| I0525 03:53:19.793846 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00463573 (* 0.0272727 = 0.000126429 loss) | |
| I0525 03:53:19.793860 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00666577 (* 0.0272727 = 0.000181794 loss) | |
| I0525 03:53:19.793874 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00297897 (* 0.0272727 = 8.12446e-05 loss) | |
| I0525 03:53:19.793889 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00544555 (* 0.0272727 = 0.000148515 loss) | |
| I0525 03:53:19.793903 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0024178 (* 0.0272727 = 6.594e-05 loss) | |
| I0525 03:53:19.793917 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00335889 (* 0.0272727 = 9.16062e-05 loss) | |
| I0525 03:53:19.793938 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0020405 (* 0.0272727 = 5.56501e-05 loss) | |
| I0525 03:53:19.793963 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0930233 | |
| I0525 03:53:19.793982 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 03:53:19.793994 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 03:53:19.794006 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 03:53:19.794018 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 03:53:19.794030 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 03:53:19.794041 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 03:53:19.794054 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 03:53:19.794066 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 03:53:19.794077 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:53:19.794090 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:53:19.794101 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:53:19.794109 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:53:19.794117 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:53:19.794129 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:53:19.794140 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:53:19.794169 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:53:19.794183 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:53:19.794194 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:53:19.794206 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:53:19.794217 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:53:19.794229 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:53:19.794240 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:53:19.794252 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.767045 | |
| I0525 03:53:19.794265 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.27907 | |
| I0525 03:53:19.794278 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.80016 (* 1 = 2.80016 loss) | |
| I0525 03:53:19.794292 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.748355 (* 1 = 0.748355 loss) | |
| I0525 03:53:19.794306 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.70138 (* 0.0909091 = 0.24558 loss) | |
| I0525 03:53:19.794320 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.99376 (* 0.0909091 = 0.27216 loss) | |
| I0525 03:53:19.794334 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.15709 (* 0.0909091 = 0.287008 loss) | |
| I0525 03:53:19.794348 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.61033 (* 0.0909091 = 0.237303 loss) | |
| I0525 03:53:19.794361 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.09067 (* 0.0909091 = 0.190061 loss) | |
| I0525 03:53:19.794375 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.76544 (* 0.0909091 = 0.160495 loss) | |
| I0525 03:53:19.794389 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.0783 (* 0.0909091 = 0.0980269 loss) | |
| I0525 03:53:19.794402 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.678254 (* 0.0909091 = 0.0616594 loss) | |
| I0525 03:53:19.794416 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0016475 (* 0.0909091 = 0.000149773 loss) | |
| I0525 03:53:19.794430 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.000643156 (* 0.0909091 = 5.84687e-05 loss) | |
| I0525 03:53:19.794445 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.000470967 (* 0.0909091 = 4.28152e-05 loss) | |
| I0525 03:53:19.794458 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.000478653 (* 0.0909091 = 4.35139e-05 loss) | |
| I0525 03:53:19.794472 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.000354277 (* 0.0909091 = 3.2207e-05 loss) | |
| I0525 03:53:19.794486 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000326132 (* 0.0909091 = 2.96484e-05 loss) | |
| I0525 03:53:19.794500 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000401394 (* 0.0909091 = 3.64904e-05 loss) | |
| I0525 03:53:19.794514 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000336714 (* 0.0909091 = 3.06103e-05 loss) | |
| I0525 03:53:19.794528 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000439536 (* 0.0909091 = 3.99578e-05 loss) | |
| I0525 03:53:19.794543 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000441357 (* 0.0909091 = 4.01234e-05 loss) | |
| I0525 03:53:19.794556 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000709273 (* 0.0909091 = 6.44794e-05 loss) | |
| I0525 03:53:19.794570 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000464645 (* 0.0909091 = 4.22405e-05 loss) | |
| I0525 03:53:19.794584 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00056146 (* 0.0909091 = 5.10418e-05 loss) | |
| I0525 03:53:19.794598 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000466338 (* 0.0909091 = 4.23944e-05 loss) | |
| I0525 03:53:19.794610 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:53:19.794622 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:53:19.794643 5272 solver.cpp:245] Train net output #149: total_confidence = 3.20007e-05 | |
| I0525 03:53:19.794656 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.64344e-05 | |
| I0525 03:53:19.794670 5272 sgd_solver.cpp:106] Iteration 17000, lr = 0.001 | |
| I0525 03:57:04.981215 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.8268 > 30) by scale factor 0.942603 | |
| I0525 03:57:54.219537 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 38.9306 > 30) by scale factor 0.770602 | |
| I0525 03:59:44.722273 5272 solver.cpp:229] Iteration 17500, loss = 9.85525 | |
| I0525 03:59:44.722414 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0833333 | |
| I0525 03:59:44.722435 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 03:59:44.722450 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 03:59:44.722462 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.25 | |
| I0525 03:59:44.722476 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 03:59:44.722488 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 03:59:44.722501 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 03:59:44.722513 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 03:59:44.722527 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 03:59:44.722538 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 03:59:44.722551 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 03:59:44.722563 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 03:59:44.722575 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 03:59:44.722587 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 03:59:44.722599 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 03:59:44.722611 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 03:59:44.722623 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 03:59:44.722635 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 03:59:44.722646 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 03:59:44.722659 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 03:59:44.722671 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 03:59:44.722682 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 03:59:44.722694 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 03:59:44.722707 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0525 03:59:44.722718 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.229167 | |
| I0525 03:59:44.722735 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.13784 (* 0.3 = 0.941352 loss) | |
| I0525 03:59:44.722749 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.928904 (* 0.3 = 0.278671 loss) | |
| I0525 03:59:44.722764 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.84019 (* 0.0272727 = 0.0774598 loss) | |
| I0525 03:59:44.722777 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.08727 (* 0.0272727 = 0.0841983 loss) | |
| I0525 03:59:44.722791 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.30686 (* 0.0272727 = 0.090187 loss) | |
| I0525 03:59:44.722805 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.55626 (* 0.0272727 = 0.0969888 loss) | |
| I0525 03:59:44.722820 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.7269 (* 0.0272727 = 0.0743701 loss) | |
| I0525 03:59:44.722833 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.59432 (* 0.0272727 = 0.0707541 loss) | |
| I0525 03:59:44.722847 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.77876 (* 0.0272727 = 0.0485115 loss) | |
| I0525 03:59:44.722862 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.4755 (* 0.0272727 = 0.0402408 loss) | |
| I0525 03:59:44.722878 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0538676 (* 0.0272727 = 0.00146912 loss) | |
| I0525 03:59:44.722893 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0453337 (* 0.0272727 = 0.00123637 loss) | |
| I0525 03:59:44.722908 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0177974 (* 0.0272727 = 0.000485385 loss) | |
| I0525 03:59:44.722923 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0158009 (* 0.0272727 = 0.000430934 loss) | |
| I0525 03:59:44.722936 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.017356 (* 0.0272727 = 0.000473346 loss) | |
| I0525 03:59:44.722970 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00948369 (* 0.0272727 = 0.000258646 loss) | |
| I0525 03:59:44.722985 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00633728 (* 0.0272727 = 0.000172835 loss) | |
| I0525 03:59:44.723000 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00526755 (* 0.0272727 = 0.000143661 loss) | |
| I0525 03:59:44.723013 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00478485 (* 0.0272727 = 0.000130496 loss) | |
| I0525 03:59:44.723027 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00294798 (* 0.0272727 = 8.03995e-05 loss) | |
| I0525 03:59:44.723042 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00626945 (* 0.0272727 = 0.000170985 loss) | |
| I0525 03:59:44.723055 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00539131 (* 0.0272727 = 0.000147036 loss) | |
| I0525 03:59:44.723069 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00288908 (* 0.0272727 = 7.87932e-05 loss) | |
| I0525 03:59:44.723084 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0035027 (* 0.0272727 = 9.55282e-05 loss) | |
| I0525 03:59:44.723096 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.125 | |
| I0525 03:59:44.723109 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.375 | |
| I0525 03:59:44.723121 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.25 | |
| I0525 03:59:44.723132 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0525 03:59:44.723145 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 03:59:44.723156 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 03:59:44.723168 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 03:59:44.723179 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 03:59:44.723191 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 03:59:44.723204 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 03:59:44.723217 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 03:59:44.723227 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 03:59:44.723238 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 03:59:44.723250 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 03:59:44.723261 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 03:59:44.723273 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 03:59:44.723284 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 03:59:44.723296 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 03:59:44.723307 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 03:59:44.723320 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 03:59:44.723330 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 03:59:44.723342 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 03:59:44.723353 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 03:59:44.723366 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.755682 | |
| I0525 03:59:44.723376 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.270833 | |
| I0525 03:59:44.723390 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.2306 (* 0.3 = 0.969181 loss) | |
| I0525 03:59:44.723404 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.977939 (* 0.3 = 0.293382 loss) | |
| I0525 03:59:44.723418 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.78538 (* 0.0272727 = 0.0759649 loss) | |
| I0525 03:59:44.723436 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.98938 (* 0.0272727 = 0.0815286 loss) | |
| I0525 03:59:44.723461 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.18361 (* 0.0272727 = 0.0868258 loss) | |
| I0525 03:59:44.723476 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.62686 (* 0.0272727 = 0.0989143 loss) | |
| I0525 03:59:44.723490 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.38687 (* 0.0272727 = 0.0650964 loss) | |
| I0525 03:59:44.723505 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.56685 (* 0.0272727 = 0.0700051 loss) | |
| I0525 03:59:44.723518 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.99854 (* 0.0272727 = 0.0545057 loss) | |
| I0525 03:59:44.723531 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.4656 (* 0.0272727 = 0.0399708 loss) | |
| I0525 03:59:44.723546 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.027803 (* 0.0272727 = 0.000758264 loss) | |
| I0525 03:59:44.723559 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0140055 (* 0.0272727 = 0.000381968 loss) | |
| I0525 03:59:44.723573 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00899451 (* 0.0272727 = 0.000245305 loss) | |
| I0525 03:59:44.723587 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00886107 (* 0.0272727 = 0.000241666 loss) | |
| I0525 03:59:44.723601 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00504581 (* 0.0272727 = 0.000137613 loss) | |
| I0525 03:59:44.723615 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00922408 (* 0.0272727 = 0.000251566 loss) | |
| I0525 03:59:44.723628 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00454956 (* 0.0272727 = 0.000124079 loss) | |
| I0525 03:59:44.723642 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00492972 (* 0.0272727 = 0.000134447 loss) | |
| I0525 03:59:44.723656 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00588692 (* 0.0272727 = 0.000160552 loss) | |
| I0525 03:59:44.723670 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00195813 (* 0.0272727 = 5.34037e-05 loss) | |
| I0525 03:59:44.723685 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00321498 (* 0.0272727 = 8.76812e-05 loss) | |
| I0525 03:59:44.723698 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00318902 (* 0.0272727 = 8.69732e-05 loss) | |
| I0525 03:59:44.723711 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00318535 (* 0.0272727 = 8.68733e-05 loss) | |
| I0525 03:59:44.723726 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00273894 (* 0.0272727 = 7.46983e-05 loss) | |
| I0525 03:59:44.723737 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0833333 | |
| I0525 03:59:44.723749 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 03:59:44.723758 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 03:59:44.723767 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 03:59:44.723778 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 03:59:44.723790 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.625 | |
| I0525 03:59:44.723803 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 03:59:44.723814 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 03:59:44.723826 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 03:59:44.723839 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 03:59:44.723850 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 03:59:44.723861 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 03:59:44.723872 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 03:59:44.723884 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 03:59:44.723896 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 03:59:44.723907 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 03:59:44.723918 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 03:59:44.723943 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 03:59:44.723956 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 03:59:44.723968 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 03:59:44.723979 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 03:59:44.723991 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 03:59:44.724002 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 03:59:44.724015 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.744318 | |
| I0525 03:59:44.724026 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.270833 | |
| I0525 03:59:44.724041 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.91138 (* 1 = 2.91138 loss) | |
| I0525 03:59:44.724055 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.839294 (* 1 = 0.839294 loss) | |
| I0525 03:59:44.724068 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.41788 (* 0.0909091 = 0.219807 loss) | |
| I0525 03:59:44.724082 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.83776 (* 0.0909091 = 0.257979 loss) | |
| I0525 03:59:44.724097 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.8583 (* 0.0909091 = 0.259845 loss) | |
| I0525 03:59:44.724109 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.33131 (* 0.0909091 = 0.302846 loss) | |
| I0525 03:59:44.724123 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.12336 (* 0.0909091 = 0.193033 loss) | |
| I0525 03:59:44.724138 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.12498 (* 0.0909091 = 0.19318 loss) | |
| I0525 03:59:44.724151 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.77389 (* 0.0909091 = 0.161262 loss) | |
| I0525 03:59:44.724164 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.24847 (* 0.0909091 = 0.113497 loss) | |
| I0525 03:59:44.724179 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0192627 (* 0.0909091 = 0.00175116 loss) | |
| I0525 03:59:44.724192 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00625155 (* 0.0909091 = 0.000568323 loss) | |
| I0525 03:59:44.724206 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00500311 (* 0.0909091 = 0.000454828 loss) | |
| I0525 03:59:44.724220 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00277463 (* 0.0909091 = 0.000252239 loss) | |
| I0525 03:59:44.724233 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00241169 (* 0.0909091 = 0.000219244 loss) | |
| I0525 03:59:44.724248 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00187107 (* 0.0909091 = 0.000170098 loss) | |
| I0525 03:59:44.724262 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00196994 (* 0.0909091 = 0.000179085 loss) | |
| I0525 03:59:44.724277 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00126634 (* 0.0909091 = 0.000115121 loss) | |
| I0525 03:59:44.724289 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00153558 (* 0.0909091 = 0.000139599 loss) | |
| I0525 03:59:44.724304 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00130161 (* 0.0909091 = 0.000118329 loss) | |
| I0525 03:59:44.724318 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00119699 (* 0.0909091 = 0.000108817 loss) | |
| I0525 03:59:44.724331 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00118888 (* 0.0909091 = 0.00010808 loss) | |
| I0525 03:59:44.724345 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000583546 (* 0.0909091 = 5.30496e-05 loss) | |
| I0525 03:59:44.724359 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000492477 (* 0.0909091 = 4.47706e-05 loss) | |
| I0525 03:59:44.724372 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 03:59:44.724383 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 03:59:44.724395 5272 solver.cpp:245] Train net output #149: total_confidence = 4.80769e-05 | |
| I0525 03:59:44.724416 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 5.82044e-05 | |
| I0525 03:59:44.724431 5272 sgd_solver.cpp:106] Iteration 17500, lr = 0.001 | |
| I0525 04:00:55.935582 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.2513 > 30) by scale factor 0.745318 | |
| I0525 04:01:12.858304 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 37.2357 > 30) by scale factor 0.805678 | |
| I0525 04:04:26.093750 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.1223 > 30) by scale factor 0.905734 | |
| I0525 04:06:09.641572 5272 solver.cpp:229] Iteration 18000, loss = 9.86657 | |
| I0525 04:06:09.641696 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0638298 | |
| I0525 04:06:09.641717 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 04:06:09.641731 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 04:06:09.641743 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 04:06:09.641755 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 04:06:09.641768 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.25 | |
| I0525 04:06:09.641782 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 04:06:09.641793 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.75 | |
| I0525 04:06:09.641806 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 04:06:09.641819 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 04:06:09.641831 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 04:06:09.641844 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 04:06:09.641855 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 04:06:09.641867 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 04:06:09.641882 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 04:06:09.641894 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 04:06:09.641906 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 04:06:09.641918 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 04:06:09.641929 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:06:09.641942 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:06:09.641953 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:06:09.641964 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:06:09.641976 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:06:09.641988 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.744318 | |
| I0525 04:06:09.642000 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.234043 | |
| I0525 04:06:09.642016 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.34429 (* 0.3 = 1.00329 loss) | |
| I0525 04:06:09.642030 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.00033 (* 0.3 = 0.300099 loss) | |
| I0525 04:06:09.642045 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.94846 (* 0.0272727 = 0.0804126 loss) | |
| I0525 04:06:09.642060 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.66943 (* 0.0272727 = 0.100075 loss) | |
| I0525 04:06:09.642073 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.46056 (* 0.0272727 = 0.0943789 loss) | |
| I0525 04:06:09.642087 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.01885 (* 0.0272727 = 0.0823323 loss) | |
| I0525 04:06:09.642102 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.07219 (* 0.0272727 = 0.0837869 loss) | |
| I0525 04:06:09.642115 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.62537 (* 0.0272727 = 0.0988738 loss) | |
| I0525 04:06:09.642128 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.38733 (* 0.0272727 = 0.0378363 loss) | |
| I0525 04:06:09.642143 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.32023 (* 0.0272727 = 0.0360063 loss) | |
| I0525 04:06:09.642156 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.101036 (* 0.0272727 = 0.00275553 loss) | |
| I0525 04:06:09.642171 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0615007 (* 0.0272727 = 0.00167729 loss) | |
| I0525 04:06:09.642185 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0641018 (* 0.0272727 = 0.00174823 loss) | |
| I0525 04:06:09.642199 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.036175 (* 0.0272727 = 0.00098659 loss) | |
| I0525 04:06:09.642215 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0259707 (* 0.0272727 = 0.000708291 loss) | |
| I0525 04:06:09.642246 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.026647 (* 0.0272727 = 0.000726735 loss) | |
| I0525 04:06:09.642261 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0186071 (* 0.0272727 = 0.000507467 loss) | |
| I0525 04:06:09.642277 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00926692 (* 0.0272727 = 0.000252734 loss) | |
| I0525 04:06:09.642290 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00917478 (* 0.0272727 = 0.000250221 loss) | |
| I0525 04:06:09.642304 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00811545 (* 0.0272727 = 0.00022133 loss) | |
| I0525 04:06:09.642318 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00526028 (* 0.0272727 = 0.000143462 loss) | |
| I0525 04:06:09.642331 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00553701 (* 0.0272727 = 0.000151009 loss) | |
| I0525 04:06:09.642345 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00336276 (* 0.0272727 = 9.17117e-05 loss) | |
| I0525 04:06:09.642359 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00503558 (* 0.0272727 = 0.000137334 loss) | |
| I0525 04:06:09.642372 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.12766 | |
| I0525 04:06:09.642385 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 04:06:09.642396 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 04:06:09.642407 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 04:06:09.642419 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 04:06:09.642431 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 04:06:09.642442 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 04:06:09.642454 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.75 | |
| I0525 04:06:09.642467 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 04:06:09.642478 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 04:06:09.642489 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 04:06:09.642500 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 04:06:09.642511 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 04:06:09.642524 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 04:06:09.642534 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 04:06:09.642545 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 04:06:09.642557 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 04:06:09.642568 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 04:06:09.642580 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:06:09.642590 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:06:09.642601 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:06:09.642613 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:06:09.642624 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:06:09.642637 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0525 04:06:09.642647 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.319149 | |
| I0525 04:06:09.642662 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.55852 (* 0.3 = 1.06756 loss) | |
| I0525 04:06:09.642674 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.05997 (* 0.3 = 0.317992 loss) | |
| I0525 04:06:09.642688 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.43709 (* 0.0272727 = 0.0937389 loss) | |
| I0525 04:06:09.642702 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 4.03267 (* 0.0272727 = 0.109982 loss) | |
| I0525 04:06:09.642730 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.67203 (* 0.0272727 = 0.100146 loss) | |
| I0525 04:06:09.642745 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.11115 (* 0.0272727 = 0.0848497 loss) | |
| I0525 04:06:09.642760 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.353 (* 0.0272727 = 0.0914455 loss) | |
| I0525 04:06:09.642773 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 4.04124 (* 0.0272727 = 0.110216 loss) | |
| I0525 04:06:09.642786 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.41236 (* 0.0272727 = 0.0385188 loss) | |
| I0525 04:06:09.642801 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.33945 (* 0.0272727 = 0.0365306 loss) | |
| I0525 04:06:09.642812 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.076946 (* 0.0272727 = 0.00209853 loss) | |
| I0525 04:06:09.642825 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0345699 (* 0.0272727 = 0.000942816 loss) | |
| I0525 04:06:09.642839 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0216829 (* 0.0272727 = 0.000591352 loss) | |
| I0525 04:06:09.642854 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0400009 (* 0.0272727 = 0.00109093 loss) | |
| I0525 04:06:09.642868 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.0256295 (* 0.0272727 = 0.000698986 loss) | |
| I0525 04:06:09.642882 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0245151 (* 0.0272727 = 0.000668594 loss) | |
| I0525 04:06:09.642895 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0146602 (* 0.0272727 = 0.000399823 loss) | |
| I0525 04:06:09.642910 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00956319 (* 0.0272727 = 0.000260814 loss) | |
| I0525 04:06:09.642923 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00697215 (* 0.0272727 = 0.00019015 loss) | |
| I0525 04:06:09.642940 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00945512 (* 0.0272727 = 0.000257867 loss) | |
| I0525 04:06:09.642954 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0048763 (* 0.0272727 = 0.00013299 loss) | |
| I0525 04:06:09.642968 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00932681 (* 0.0272727 = 0.000254368 loss) | |
| I0525 04:06:09.642982 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00343135 (* 0.0272727 = 9.35822e-05 loss) | |
| I0525 04:06:09.642995 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00630837 (* 0.0272727 = 0.000172047 loss) | |
| I0525 04:06:09.643007 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0851064 | |
| I0525 04:06:09.643019 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 04:06:09.643031 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 04:06:09.643043 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 04:06:09.643054 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.375 | |
| I0525 04:06:09.643065 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 04:06:09.643077 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 04:06:09.643090 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 04:06:09.643100 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 04:06:09.643112 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 04:06:09.643123 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 04:06:09.643134 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 04:06:09.643146 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 04:06:09.643157 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 04:06:09.643168 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 04:06:09.643179 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 04:06:09.643190 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 04:06:09.643213 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 04:06:09.643224 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:06:09.643236 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:06:09.643247 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:06:09.643259 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:06:09.643270 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:06:09.643281 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.75 | |
| I0525 04:06:09.643293 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.212766 | |
| I0525 04:06:09.643306 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.62987 (* 1 = 3.62987 loss) | |
| I0525 04:06:09.643321 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.08267 (* 1 = 1.08267 loss) | |
| I0525 04:06:09.643334 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.01471 (* 0.0909091 = 0.274065 loss) | |
| I0525 04:06:09.643347 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 4.37728 (* 0.0909091 = 0.397935 loss) | |
| I0525 04:06:09.643362 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.62819 (* 0.0909091 = 0.329836 loss) | |
| I0525 04:06:09.643375 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.91936 (* 0.0909091 = 0.265396 loss) | |
| I0525 04:06:09.643388 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.55817 (* 0.0909091 = 0.32347 loss) | |
| I0525 04:06:09.643402 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.52315 (* 0.0909091 = 0.320287 loss) | |
| I0525 04:06:09.643415 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.21667 (* 0.0909091 = 0.110607 loss) | |
| I0525 04:06:09.643429 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.02342 (* 0.0909091 = 0.0930385 loss) | |
| I0525 04:06:09.643443 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0809822 (* 0.0909091 = 0.00736202 loss) | |
| I0525 04:06:09.643457 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0541274 (* 0.0909091 = 0.00492067 loss) | |
| I0525 04:06:09.643471 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0311399 (* 0.0909091 = 0.0028309 loss) | |
| I0525 04:06:09.643486 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0116313 (* 0.0909091 = 0.00105739 loss) | |
| I0525 04:06:09.643499 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0157709 (* 0.0909091 = 0.00143371 loss) | |
| I0525 04:06:09.643513 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00845612 (* 0.0909091 = 0.000768738 loss) | |
| I0525 04:06:09.643528 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00931627 (* 0.0909091 = 0.000846933 loss) | |
| I0525 04:06:09.643543 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00612826 (* 0.0909091 = 0.000557114 loss) | |
| I0525 04:06:09.643556 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00472019 (* 0.0909091 = 0.000429108 loss) | |
| I0525 04:06:09.643569 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00234181 (* 0.0909091 = 0.000212892 loss) | |
| I0525 04:06:09.643584 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00185075 (* 0.0909091 = 0.00016825 loss) | |
| I0525 04:06:09.643597 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00208436 (* 0.0909091 = 0.000189487 loss) | |
| I0525 04:06:09.643610 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000702511 (* 0.0909091 = 6.38647e-05 loss) | |
| I0525 04:06:09.643625 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000892533 (* 0.0909091 = 8.11394e-05 loss) | |
| I0525 04:06:09.643636 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:06:09.643647 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:06:09.643658 5272 solver.cpp:245] Train net output #149: total_confidence = 0.000318692 | |
| I0525 04:06:09.643679 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00203058 | |
| I0525 04:06:09.643693 5272 sgd_solver.cpp:106] Iteration 18000, lr = 0.001 | |
| I0525 04:06:53.103301 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.3493 > 30) by scale factor 0.98849 | |
| I0525 04:12:34.594841 5272 solver.cpp:229] Iteration 18500, loss = 9.77522 | |
| I0525 04:12:34.594987 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.08 | |
| I0525 04:12:34.595008 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 04:12:34.595021 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 04:12:34.595034 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 04:12:34.595049 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 04:12:34.595062 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 04:12:34.595074 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 04:12:34.595088 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 04:12:34.595099 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 04:12:34.595113 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0525 04:12:34.595124 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 04:12:34.595137 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 04:12:34.595151 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 04:12:34.595163 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 04:12:34.595175 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 04:12:34.595186 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 04:12:34.595198 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 04:12:34.595211 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 04:12:34.595222 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:12:34.595234 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:12:34.595247 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:12:34.595257 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:12:34.595269 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:12:34.595281 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.732955 | |
| I0525 04:12:34.595293 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.34 | |
| I0525 04:12:34.595309 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.16741 (* 0.3 = 0.950222 loss) | |
| I0525 04:12:34.595324 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.02969 (* 0.3 = 0.308907 loss) | |
| I0525 04:12:34.595338 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.03737 (* 0.0272727 = 0.0828374 loss) | |
| I0525 04:12:34.595352 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.51136 (* 0.0272727 = 0.0957642 loss) | |
| I0525 04:12:34.595366 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.32134 (* 0.0272727 = 0.0905819 loss) | |
| I0525 04:12:34.595381 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.22547 (* 0.0272727 = 0.0879675 loss) | |
| I0525 04:12:34.595394 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.3253 (* 0.0272727 = 0.0634173 loss) | |
| I0525 04:12:34.595408 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.83675 (* 0.0272727 = 0.0500933 loss) | |
| I0525 04:12:34.595422 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.86559 (* 0.0272727 = 0.0508797 loss) | |
| I0525 04:12:34.595437 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.16717 (* 0.0272727 = 0.0318318 loss) | |
| I0525 04:12:34.595450 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.04078 (* 0.0272727 = 0.0283849 loss) | |
| I0525 04:12:34.595465 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.530819 (* 0.0272727 = 0.0144769 loss) | |
| I0525 04:12:34.595479 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.525478 (* 0.0272727 = 0.0143312 loss) | |
| I0525 04:12:34.595494 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.158687 (* 0.0272727 = 0.00432782 loss) | |
| I0525 04:12:34.595509 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0694457 (* 0.0272727 = 0.00189397 loss) | |
| I0525 04:12:34.595543 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0769427 (* 0.0272727 = 0.00209844 loss) | |
| I0525 04:12:34.595559 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0511271 (* 0.0272727 = 0.00139438 loss) | |
| I0525 04:12:34.595574 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0280499 (* 0.0272727 = 0.000764996 loss) | |
| I0525 04:12:34.595588 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0122903 (* 0.0272727 = 0.000335189 loss) | |
| I0525 04:12:34.595602 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0102983 (* 0.0272727 = 0.000280864 loss) | |
| I0525 04:12:34.595616 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00747338 (* 0.0272727 = 0.000203819 loss) | |
| I0525 04:12:34.595630 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00775103 (* 0.0272727 = 0.000211392 loss) | |
| I0525 04:12:34.595643 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00526831 (* 0.0272727 = 0.000143681 loss) | |
| I0525 04:12:34.595657 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0128368 (* 0.0272727 = 0.000350095 loss) | |
| I0525 04:12:34.595670 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.18 | |
| I0525 04:12:34.595682 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.25 | |
| I0525 04:12:34.595695 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 04:12:34.595706 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 04:12:34.595718 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.375 | |
| I0525 04:12:34.595731 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 04:12:34.595742 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.5 | |
| I0525 04:12:34.595754 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 04:12:34.595767 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 04:12:34.595778 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0525 04:12:34.595790 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 04:12:34.595803 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 04:12:34.595813 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 04:12:34.595825 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 04:12:34.595837 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 04:12:34.595849 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 04:12:34.595860 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 04:12:34.595872 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 04:12:34.595887 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:12:34.595899 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:12:34.595911 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:12:34.595923 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:12:34.595935 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:12:34.595947 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.75 | |
| I0525 04:12:34.595958 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.28 | |
| I0525 04:12:34.595973 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.03086 (* 0.3 = 0.909258 loss) | |
| I0525 04:12:34.595986 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.03762 (* 0.3 = 0.311286 loss) | |
| I0525 04:12:34.596006 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.02777 (* 0.0272727 = 0.0825757 loss) | |
| I0525 04:12:34.596034 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.12939 (* 0.0272727 = 0.0853469 loss) | |
| I0525 04:12:34.596067 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.26086 (* 0.0272727 = 0.0889324 loss) | |
| I0525 04:12:34.596083 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.37048 (* 0.0272727 = 0.0919223 loss) | |
| I0525 04:12:34.596098 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.03878 (* 0.0272727 = 0.055603 loss) | |
| I0525 04:12:34.596112 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 1.70372 (* 0.0272727 = 0.046465 loss) | |
| I0525 04:12:34.596125 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.70922 (* 0.0272727 = 0.0466152 loss) | |
| I0525 04:12:34.596139 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.17508 (* 0.0272727 = 0.0320476 loss) | |
| I0525 04:12:34.596153 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.909057 (* 0.0272727 = 0.0247925 loss) | |
| I0525 04:12:34.596168 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.611669 (* 0.0272727 = 0.0166819 loss) | |
| I0525 04:12:34.596180 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.603757 (* 0.0272727 = 0.0164661 loss) | |
| I0525 04:12:34.596195 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.208195 (* 0.0272727 = 0.00567804 loss) | |
| I0525 04:12:34.596210 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.181189 (* 0.0272727 = 0.00494151 loss) | |
| I0525 04:12:34.596223 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.149976 (* 0.0272727 = 0.00409025 loss) | |
| I0525 04:12:34.596236 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0940518 (* 0.0272727 = 0.00256505 loss) | |
| I0525 04:12:34.596251 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0471108 (* 0.0272727 = 0.00128484 loss) | |
| I0525 04:12:34.596264 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0277911 (* 0.0272727 = 0.000757939 loss) | |
| I0525 04:12:34.596278 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0206049 (* 0.0272727 = 0.000561952 loss) | |
| I0525 04:12:34.596292 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.028789 (* 0.0272727 = 0.000785153 loss) | |
| I0525 04:12:34.596307 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.0108661 (* 0.0272727 = 0.000296348 loss) | |
| I0525 04:12:34.596319 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0216374 (* 0.0272727 = 0.000590111 loss) | |
| I0525 04:12:34.596333 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0180423 (* 0.0272727 = 0.000492064 loss) | |
| I0525 04:12:34.596346 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.14 | |
| I0525 04:12:34.596359 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.375 | |
| I0525 04:12:34.596370 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 04:12:34.596382 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 04:12:34.596395 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 04:12:34.596406 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.5 | |
| I0525 04:12:34.596418 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 04:12:34.596431 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 04:12:34.596442 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 04:12:34.596453 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0525 04:12:34.596465 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 04:12:34.596477 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 04:12:34.596489 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 04:12:34.596501 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 04:12:34.596513 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 04:12:34.596524 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 04:12:34.596535 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 04:12:34.596556 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 04:12:34.596570 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:12:34.596580 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:12:34.596592 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:12:34.596604 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:12:34.596616 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:12:34.596627 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.732955 | |
| I0525 04:12:34.596639 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.34 | |
| I0525 04:12:34.596662 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.0144 (* 1 = 3.0144 loss) | |
| I0525 04:12:34.596690 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.03708 (* 1 = 1.03708 loss) | |
| I0525 04:12:34.596709 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.70461 (* 0.0909091 = 0.245874 loss) | |
| I0525 04:12:34.596719 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.05563 (* 0.0909091 = 0.277785 loss) | |
| I0525 04:12:34.596729 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.28489 (* 0.0909091 = 0.298627 loss) | |
| I0525 04:12:34.596743 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.98885 (* 0.0909091 = 0.271714 loss) | |
| I0525 04:12:34.596757 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 1.91774 (* 0.0909091 = 0.17434 loss) | |
| I0525 04:12:34.596771 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.77225 (* 0.0909091 = 0.161113 loss) | |
| I0525 04:12:34.596784 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.43922 (* 0.0909091 = 0.130838 loss) | |
| I0525 04:12:34.596798 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.94987 (* 0.0909091 = 0.0863518 loss) | |
| I0525 04:12:34.596812 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.91769 (* 0.0909091 = 0.0834264 loss) | |
| I0525 04:12:34.596825 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.490146 (* 0.0909091 = 0.0445587 loss) | |
| I0525 04:12:34.596839 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.484407 (* 0.0909091 = 0.044037 loss) | |
| I0525 04:12:34.596853 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.128122 (* 0.0909091 = 0.0116474 loss) | |
| I0525 04:12:34.596868 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0959597 (* 0.0909091 = 0.00872361 loss) | |
| I0525 04:12:34.596881 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0480195 (* 0.0909091 = 0.00436541 loss) | |
| I0525 04:12:34.596895 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0211426 (* 0.0909091 = 0.00192205 loss) | |
| I0525 04:12:34.596909 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0135416 (* 0.0909091 = 0.00123106 loss) | |
| I0525 04:12:34.596923 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.013089 (* 0.0909091 = 0.00118991 loss) | |
| I0525 04:12:34.596941 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00712528 (* 0.0909091 = 0.000647753 loss) | |
| I0525 04:12:34.596956 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00629259 (* 0.0909091 = 0.000572054 loss) | |
| I0525 04:12:34.596969 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00380776 (* 0.0909091 = 0.00034616 loss) | |
| I0525 04:12:34.596982 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00170268 (* 0.0909091 = 0.000154789 loss) | |
| I0525 04:12:34.596997 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00138994 (* 0.0909091 = 0.000126358 loss) | |
| I0525 04:12:34.597008 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:12:34.597019 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:12:34.597031 5272 solver.cpp:245] Train net output #149: total_confidence = 5.47178e-05 | |
| I0525 04:12:34.597057 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00158301 | |
| I0525 04:12:34.597072 5272 sgd_solver.cpp:106] Iteration 18500, lr = 0.001 | |
| I0525 04:16:39.772830 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.709 > 30) by scale factor 0.976911 | |
| I0525 04:16:55.974751 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.0794 > 30) by scale factor 0.96527 | |
| I0525 04:18:59.588925 5272 solver.cpp:229] Iteration 19000, loss = 9.86371 | |
| I0525 04:18:59.589085 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0882353 | |
| I0525 04:18:59.589108 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 04:18:59.589121 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 04:18:59.589134 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 04:18:59.589146 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 04:18:59.589159 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 04:18:59.589171 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.125 | |
| I0525 04:18:59.589184 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 04:18:59.589195 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.625 | |
| I0525 04:18:59.589223 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.625 | |
| I0525 04:18:59.589236 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.625 | |
| I0525 04:18:59.589249 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 04:18:59.589262 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 04:18:59.589274 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 04:18:59.589287 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 04:18:59.589299 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 04:18:59.589311 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 04:18:59.589323 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 04:18:59.589335 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:18:59.589347 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:18:59.589359 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:18:59.589371 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:18:59.589382 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:18:59.589395 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.642045 | |
| I0525 04:18:59.589406 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.25 | |
| I0525 04:18:59.589423 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.26631 (* 0.3 = 0.979893 loss) | |
| I0525 04:18:59.589437 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.43865 (* 0.3 = 0.431595 loss) | |
| I0525 04:18:59.589452 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.19737 (* 0.0272727 = 0.087201 loss) | |
| I0525 04:18:59.589467 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.39988 (* 0.0272727 = 0.0927241 loss) | |
| I0525 04:18:59.589480 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.79299 (* 0.0272727 = 0.103445 loss) | |
| I0525 04:18:59.589494 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.91754 (* 0.0272727 = 0.106842 loss) | |
| I0525 04:18:59.589509 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.15509 (* 0.0272727 = 0.0860479 loss) | |
| I0525 04:18:59.589522 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 4.21535 (* 0.0272727 = 0.114964 loss) | |
| I0525 04:18:59.589536 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.62505 (* 0.0272727 = 0.0715922 loss) | |
| I0525 04:18:59.589550 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.62388 (* 0.0272727 = 0.0442875 loss) | |
| I0525 04:18:59.589565 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.40825 (* 0.0272727 = 0.0384069 loss) | |
| I0525 04:18:59.589579 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 1.38675 (* 0.0272727 = 0.0378204 loss) | |
| I0525 04:18:59.589593 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.643883 (* 0.0272727 = 0.0175605 loss) | |
| I0525 04:18:59.589607 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.922386 (* 0.0272727 = 0.025156 loss) | |
| I0525 04:18:59.589622 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.480812 (* 0.0272727 = 0.0131131 loss) | |
| I0525 04:18:59.589658 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.696563 (* 0.0272727 = 0.0189972 loss) | |
| I0525 04:18:59.589674 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.748137 (* 0.0272727 = 0.0204037 loss) | |
| I0525 04:18:59.589689 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.834522 (* 0.0272727 = 0.0227597 loss) | |
| I0525 04:18:59.589704 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.031992 (* 0.0272727 = 0.000872508 loss) | |
| I0525 04:18:59.589717 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0180784 (* 0.0272727 = 0.000493046 loss) | |
| I0525 04:18:59.589732 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0155258 (* 0.0272727 = 0.00042343 loss) | |
| I0525 04:18:59.589746 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0126712 (* 0.0272727 = 0.000345578 loss) | |
| I0525 04:18:59.589761 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0125072 (* 0.0272727 = 0.000341105 loss) | |
| I0525 04:18:59.589774 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0190095 (* 0.0272727 = 0.000518441 loss) | |
| I0525 04:18:59.589787 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0441176 | |
| I0525 04:18:59.589799 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 04:18:59.589812 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 04:18:59.589824 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 04:18:59.589836 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 04:18:59.589848 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 04:18:59.589859 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 04:18:59.589871 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 04:18:59.589886 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.625 | |
| I0525 04:18:59.589898 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.625 | |
| I0525 04:18:59.589911 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.625 | |
| I0525 04:18:59.589922 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 04:18:59.589934 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 04:18:59.589946 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 04:18:59.589958 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 04:18:59.589970 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 04:18:59.589982 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 04:18:59.589994 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 04:18:59.590006 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:18:59.590018 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:18:59.590029 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:18:59.590040 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:18:59.590052 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:18:59.590064 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.630682 | |
| I0525 04:18:59.590075 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0525 04:18:59.590092 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.22183 (* 0.3 = 0.966549 loss) | |
| I0525 04:18:59.590107 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.35254 (* 0.3 = 0.405762 loss) | |
| I0525 04:18:59.590121 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.11079 (* 0.0272727 = 0.0848398 loss) | |
| I0525 04:18:59.590136 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.2504 (* 0.0272727 = 0.0886473 loss) | |
| I0525 04:18:59.590159 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.51762 (* 0.0272727 = 0.095935 loss) | |
| I0525 04:18:59.590174 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.60495 (* 0.0272727 = 0.0983167 loss) | |
| I0525 04:18:59.590188 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.54035 (* 0.0272727 = 0.096555 loss) | |
| I0525 04:18:59.590203 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.58681 (* 0.0272727 = 0.0978221 loss) | |
| I0525 04:18:59.590215 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.61884 (* 0.0272727 = 0.0714229 loss) | |
| I0525 04:18:59.590229 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.66121 (* 0.0272727 = 0.0453056 loss) | |
| I0525 04:18:59.590243 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.31875 (* 0.0272727 = 0.0359659 loss) | |
| I0525 04:18:59.590257 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 1.51599 (* 0.0272727 = 0.0413452 loss) | |
| I0525 04:18:59.590270 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.628763 (* 0.0272727 = 0.0171481 loss) | |
| I0525 04:18:59.590284 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.654404 (* 0.0272727 = 0.0178474 loss) | |
| I0525 04:18:59.590298 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.492787 (* 0.0272727 = 0.0134397 loss) | |
| I0525 04:18:59.590312 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.786999 (* 0.0272727 = 0.0214636 loss) | |
| I0525 04:18:59.590327 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.656859 (* 0.0272727 = 0.0179143 loss) | |
| I0525 04:18:59.590340 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.776949 (* 0.0272727 = 0.0211895 loss) | |
| I0525 04:18:59.590354 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0135091 (* 0.0272727 = 0.000368431 loss) | |
| I0525 04:18:59.590368 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0080268 (* 0.0272727 = 0.000218913 loss) | |
| I0525 04:18:59.590383 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00940633 (* 0.0272727 = 0.000256536 loss) | |
| I0525 04:18:59.590396 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00729824 (* 0.0272727 = 0.000199043 loss) | |
| I0525 04:18:59.590410 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00494928 (* 0.0272727 = 0.00013498 loss) | |
| I0525 04:18:59.590425 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00607485 (* 0.0272727 = 0.000165678 loss) | |
| I0525 04:18:59.590436 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.117647 | |
| I0525 04:18:59.590448 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.25 | |
| I0525 04:18:59.590463 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 04:18:59.590476 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 04:18:59.590487 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 04:18:59.590498 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 04:18:59.590510 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.125 | |
| I0525 04:18:59.590523 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.375 | |
| I0525 04:18:59.590534 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.625 | |
| I0525 04:18:59.590546 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0525 04:18:59.590558 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.625 | |
| I0525 04:18:59.590569 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 04:18:59.590581 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 04:18:59.590593 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 04:18:59.590605 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 04:18:59.590616 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 04:18:59.590628 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 04:18:59.590651 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 04:18:59.590663 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:18:59.590674 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:18:59.590687 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:18:59.590698 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:18:59.590710 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:18:59.590721 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.659091 | |
| I0525 04:18:59.590734 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.264706 | |
| I0525 04:18:59.590747 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.17467 (* 1 = 3.17467 loss) | |
| I0525 04:18:59.590761 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.3034 (* 1 = 1.3034 loss) | |
| I0525 04:18:59.590775 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.85527 (* 0.0909091 = 0.25957 loss) | |
| I0525 04:18:59.590790 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.7749 (* 0.0909091 = 0.252264 loss) | |
| I0525 04:18:59.590802 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.25527 (* 0.0909091 = 0.295934 loss) | |
| I0525 04:18:59.590816 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.24903 (* 0.0909091 = 0.295366 loss) | |
| I0525 04:18:59.590831 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.29476 (* 0.0909091 = 0.299524 loss) | |
| I0525 04:18:59.590844 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.48059 (* 0.0909091 = 0.316417 loss) | |
| I0525 04:18:59.590857 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.6687 (* 0.0909091 = 0.242609 loss) | |
| I0525 04:18:59.590872 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.58777 (* 0.0909091 = 0.144343 loss) | |
| I0525 04:18:59.590885 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 1.27067 (* 0.0909091 = 0.115516 loss) | |
| I0525 04:18:59.590899 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 1.29685 (* 0.0909091 = 0.117895 loss) | |
| I0525 04:18:59.590912 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.493306 (* 0.0909091 = 0.044846 loss) | |
| I0525 04:18:59.590929 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.550736 (* 0.0909091 = 0.0500669 loss) | |
| I0525 04:18:59.590944 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.403936 (* 0.0909091 = 0.0367215 loss) | |
| I0525 04:18:59.590957 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.686864 (* 0.0909091 = 0.0624422 loss) | |
| I0525 04:18:59.590971 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.638432 (* 0.0909091 = 0.0580392 loss) | |
| I0525 04:18:59.590984 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.57413 (* 0.0909091 = 0.0521936 loss) | |
| I0525 04:18:59.590999 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0270174 (* 0.0909091 = 0.00245612 loss) | |
| I0525 04:18:59.591012 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0192084 (* 0.0909091 = 0.00174621 loss) | |
| I0525 04:18:59.591027 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.0116814 (* 0.0909091 = 0.00106195 loss) | |
| I0525 04:18:59.591040 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00527394 (* 0.0909091 = 0.000479449 loss) | |
| I0525 04:18:59.591054 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00356909 (* 0.0909091 = 0.000324463 loss) | |
| I0525 04:18:59.591068 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00289446 (* 0.0909091 = 0.000263133 loss) | |
| I0525 04:18:59.591080 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:18:59.591092 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:18:59.591104 5272 solver.cpp:245] Train net output #149: total_confidence = 2.56861e-07 | |
| I0525 04:18:59.591125 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 1.77086e-06 | |
| I0525 04:18:59.591140 5272 sgd_solver.cpp:106] Iteration 19000, lr = 0.001 | |
| I0525 04:22:20.113984 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 45.9868 > 30) by scale factor 0.652361 | |
| I0525 04:23:02.446341 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.603 > 30) by scale factor 0.892777 | |
| I0525 04:24:05.590708 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.2828 > 30) by scale factor 0.929286 | |
| I0525 04:24:45.609282 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.6862 > 30) by scale factor 0.977639 | |
| I0525 04:25:24.513455 5272 solver.cpp:229] Iteration 19500, loss = 9.75068 | |
| I0525 04:25:24.513592 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0816327 | |
| I0525 04:25:24.513613 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 04:25:24.513628 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.25 | |
| I0525 04:25:24.513640 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 04:25:24.513653 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.125 | |
| I0525 04:25:24.513666 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 04:25:24.513679 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 04:25:24.513691 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 04:25:24.513705 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 04:25:24.513717 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 04:25:24.513731 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 04:25:24.513742 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 04:25:24.513754 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 04:25:24.513767 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 04:25:24.513778 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 04:25:24.513790 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 04:25:24.513803 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 04:25:24.513814 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 04:25:24.513825 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:25:24.513839 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:25:24.513850 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:25:24.513862 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:25:24.513877 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:25:24.513890 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.738636 | |
| I0525 04:25:24.513902 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.22449 | |
| I0525 04:25:24.513918 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.22795 (* 0.3 = 0.968385 loss) | |
| I0525 04:25:24.513933 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.942298 (* 0.3 = 0.282689 loss) | |
| I0525 04:25:24.513948 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.39759 (* 0.0272727 = 0.0926616 loss) | |
| I0525 04:25:24.513962 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.30668 (* 0.0272727 = 0.0901821 loss) | |
| I0525 04:25:24.513977 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.23659 (* 0.0272727 = 0.0882707 loss) | |
| I0525 04:25:24.513990 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.23016 (* 0.0272727 = 0.0880954 loss) | |
| I0525 04:25:24.514004 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.29154 (* 0.0272727 = 0.0897692 loss) | |
| I0525 04:25:24.514019 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.50149 (* 0.0272727 = 0.0954953 loss) | |
| I0525 04:25:24.514032 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.29876 (* 0.0272727 = 0.0354207 loss) | |
| I0525 04:25:24.514046 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.896999 (* 0.0272727 = 0.0244636 loss) | |
| I0525 04:25:24.514061 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.197984 (* 0.0272727 = 0.00539957 loss) | |
| I0525 04:25:24.514075 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.116555 (* 0.0272727 = 0.00317878 loss) | |
| I0525 04:25:24.514089 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0498358 (* 0.0272727 = 0.00135916 loss) | |
| I0525 04:25:24.514104 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0660184 (* 0.0272727 = 0.0018005 loss) | |
| I0525 04:25:24.514118 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.05327 (* 0.0272727 = 0.00145282 loss) | |
| I0525 04:25:24.514153 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0359881 (* 0.0272727 = 0.000981494 loss) | |
| I0525 04:25:24.514168 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0251916 (* 0.0272727 = 0.000687044 loss) | |
| I0525 04:25:24.514183 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.031323 (* 0.0272727 = 0.000854265 loss) | |
| I0525 04:25:24.514196 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0154739 (* 0.0272727 = 0.000422016 loss) | |
| I0525 04:25:24.514210 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0133223 (* 0.0272727 = 0.000363335 loss) | |
| I0525 04:25:24.514225 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0105633 (* 0.0272727 = 0.000288091 loss) | |
| I0525 04:25:24.514238 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0106521 (* 0.0272727 = 0.000290512 loss) | |
| I0525 04:25:24.514252 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0191888 (* 0.0272727 = 0.000523332 loss) | |
| I0525 04:25:24.514266 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00957829 (* 0.0272727 = 0.000261226 loss) | |
| I0525 04:25:24.514279 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0408163 | |
| I0525 04:25:24.514292 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 04:25:24.514303 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 04:25:24.514315 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.375 | |
| I0525 04:25:24.514328 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 04:25:24.514338 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 04:25:24.514350 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 04:25:24.514363 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 04:25:24.514374 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 04:25:24.514386 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 04:25:24.514397 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 04:25:24.514408 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 04:25:24.514420 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 04:25:24.514431 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 04:25:24.514444 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 04:25:24.514456 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 04:25:24.514467 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 04:25:24.514478 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 04:25:24.514489 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:25:24.514502 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:25:24.514513 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:25:24.514525 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:25:24.514536 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:25:24.514549 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.727273 | |
| I0525 04:25:24.514560 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.22449 | |
| I0525 04:25:24.514575 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.06843 (* 0.3 = 0.92053 loss) | |
| I0525 04:25:24.514588 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.9118 (* 0.3 = 0.27354 loss) | |
| I0525 04:25:24.514605 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.7041 (* 0.0272727 = 0.0737483 loss) | |
| I0525 04:25:24.514619 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.80397 (* 0.0272727 = 0.0764719 loss) | |
| I0525 04:25:24.514643 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 2.85946 (* 0.0272727 = 0.0779854 loss) | |
| I0525 04:25:24.514658 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.27302 (* 0.0272727 = 0.0892643 loss) | |
| I0525 04:25:24.514672 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.38045 (* 0.0272727 = 0.0921942 loss) | |
| I0525 04:25:24.514686 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.34108 (* 0.0272727 = 0.0911203 loss) | |
| I0525 04:25:24.514700 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 1.46775 (* 0.0272727 = 0.0400297 loss) | |
| I0525 04:25:24.514714 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.05023 (* 0.0272727 = 0.0286426 loss) | |
| I0525 04:25:24.514729 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.114665 (* 0.0272727 = 0.00312723 loss) | |
| I0525 04:25:24.514742 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0252047 (* 0.0272727 = 0.0006874 loss) | |
| I0525 04:25:24.514756 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0176146 (* 0.0272727 = 0.000480398 loss) | |
| I0525 04:25:24.514770 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0153371 (* 0.0272727 = 0.000418285 loss) | |
| I0525 04:25:24.514785 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00847213 (* 0.0272727 = 0.000231058 loss) | |
| I0525 04:25:24.514798 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.0047955 (* 0.0272727 = 0.000130786 loss) | |
| I0525 04:25:24.514812 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00719247 (* 0.0272727 = 0.000196158 loss) | |
| I0525 04:25:24.514827 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0040245 (* 0.0272727 = 0.000109759 loss) | |
| I0525 04:25:24.514840 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00195702 (* 0.0272727 = 5.33734e-05 loss) | |
| I0525 04:25:24.514854 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00115778 (* 0.0272727 = 3.15758e-05 loss) | |
| I0525 04:25:24.514868 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0013391 (* 0.0272727 = 3.65209e-05 loss) | |
| I0525 04:25:24.514883 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.000751598 (* 0.0272727 = 2.04981e-05 loss) | |
| I0525 04:25:24.514896 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.000620915 (* 0.0272727 = 1.6934e-05 loss) | |
| I0525 04:25:24.514910 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.000676594 (* 0.0272727 = 1.84526e-05 loss) | |
| I0525 04:25:24.514922 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.163265 | |
| I0525 04:25:24.514937 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 04:25:24.514950 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 04:25:24.514961 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 04:25:24.514973 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 04:25:24.514984 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 04:25:24.514997 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 04:25:24.515008 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.75 | |
| I0525 04:25:24.515020 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 04:25:24.515031 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 04:25:24.515043 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 04:25:24.515055 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 04:25:24.515066 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 04:25:24.515079 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 04:25:24.515090 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 04:25:24.515099 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 04:25:24.515106 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 04:25:24.515127 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 04:25:24.515141 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:25:24.515152 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:25:24.515163 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:25:24.515175 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:25:24.515187 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:25:24.515198 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.761364 | |
| I0525 04:25:24.515210 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.285714 | |
| I0525 04:25:24.515224 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.92841 (* 1 = 2.92841 loss) | |
| I0525 04:25:24.515239 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.885311 (* 1 = 0.885311 loss) | |
| I0525 04:25:24.515252 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.77741 (* 0.0909091 = 0.252492 loss) | |
| I0525 04:25:24.515266 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.75129 (* 0.0909091 = 0.250117 loss) | |
| I0525 04:25:24.515280 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.8153 (* 0.0909091 = 0.255936 loss) | |
| I0525 04:25:24.515295 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.98828 (* 0.0909091 = 0.271662 loss) | |
| I0525 04:25:24.515308 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.08261 (* 0.0909091 = 0.280237 loss) | |
| I0525 04:25:24.515321 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.05782 (* 0.0909091 = 0.277984 loss) | |
| I0525 04:25:24.515336 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.12113 (* 0.0909091 = 0.101921 loss) | |
| I0525 04:25:24.515349 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.901134 (* 0.0909091 = 0.0819213 loss) | |
| I0525 04:25:24.515363 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0784026 (* 0.0909091 = 0.00712751 loss) | |
| I0525 04:25:24.515377 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0403155 (* 0.0909091 = 0.00366504 loss) | |
| I0525 04:25:24.515391 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.0103662 (* 0.0909091 = 0.000942383 loss) | |
| I0525 04:25:24.515405 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00974966 (* 0.0909091 = 0.000886333 loss) | |
| I0525 04:25:24.515419 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.0152279 (* 0.0909091 = 0.00138436 loss) | |
| I0525 04:25:24.515432 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.0107611 (* 0.0909091 = 0.000978286 loss) | |
| I0525 04:25:24.515446 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0100758 (* 0.0909091 = 0.000915979 loss) | |
| I0525 04:25:24.515460 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00648123 (* 0.0909091 = 0.000589203 loss) | |
| I0525 04:25:24.515473 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00229293 (* 0.0909091 = 0.000208448 loss) | |
| I0525 04:25:24.515487 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00254153 (* 0.0909091 = 0.000231048 loss) | |
| I0525 04:25:24.515501 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00149377 (* 0.0909091 = 0.000135797 loss) | |
| I0525 04:25:24.515514 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00116375 (* 0.0909091 = 0.000105796 loss) | |
| I0525 04:25:24.515528 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000750387 (* 0.0909091 = 6.8217e-05 loss) | |
| I0525 04:25:24.515542 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000870525 (* 0.0909091 = 7.91386e-05 loss) | |
| I0525 04:25:24.515554 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:25:24.515566 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:25:24.515578 5272 solver.cpp:245] Train net output #149: total_confidence = 3.86508e-05 | |
| I0525 04:25:24.515599 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000312713 | |
| I0525 04:25:24.515612 5272 sgd_solver.cpp:106] Iteration 19500, lr = 0.001 | |
| I0525 04:29:15.156944 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.1318 > 30) by scale factor 0.995625 | |
| I0525 04:31:49.165145 5272 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm20_iter_20000.caffemodel | |
| I0525 04:31:50.776320 5272 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm20_iter_20000.solverstate | |
| I0525 04:31:51.056249 5272 solver.cpp:338] Iteration 20000, Testing net (#0) | |
| I0525 04:32:49.085144 5272 solver.cpp:393] Test loss: 9.06269 | |
| I0525 04:32:49.085259 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0681274 | |
| I0525 04:32:49.085279 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.146 | |
| I0525 04:32:49.085294 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.105 | |
| I0525 04:32:49.085307 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.09 | |
| I0525 04:32:49.085319 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.172 | |
| I0525 04:32:49.085331 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.334 | |
| I0525 04:32:49.085343 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.476 | |
| I0525 04:32:49.085355 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.737 | |
| I0525 04:32:49.085367 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.917 | |
| I0525 04:32:49.085381 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.98 | |
| I0525 04:32:49.085392 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.993 | |
| I0525 04:32:49.085403 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 1 | |
| I0525 04:32:49.085415 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 1 | |
| I0525 04:32:49.085427 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 1 | |
| I0525 04:32:49.085438 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 1 | |
| I0525 04:32:49.085449 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 1 | |
| I0525 04:32:49.085461 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 1 | |
| I0525 04:32:49.085472 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 1 | |
| I0525 04:32:49.085484 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 1 | |
| I0525 04:32:49.085494 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 1 | |
| I0525 04:32:49.085506 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0525 04:32:49.085517 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0525 04:32:49.085530 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0525 04:32:49.085541 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.766455 | |
| I0525 04:32:49.085552 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.239561 | |
| I0525 04:32:49.085567 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.38716 (* 0.3 = 1.01615 loss) | |
| I0525 04:32:49.085582 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.878778 (* 0.3 = 0.263634 loss) | |
| I0525 04:32:49.085597 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 2.8544 (* 0.0272727 = 0.0778471 loss) | |
| I0525 04:32:49.085610 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.08968 (* 0.0272727 = 0.084264 loss) | |
| I0525 04:32:49.085623 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.16894 (* 0.0272727 = 0.0864256 loss) | |
| I0525 04:32:49.085638 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 3.01095 (* 0.0272727 = 0.0821167 loss) | |
| I0525 04:32:49.085650 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 2.51506 (* 0.0272727 = 0.0685926 loss) | |
| I0525 04:32:49.085664 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.0789 (* 0.0272727 = 0.0566972 loss) | |
| I0525 04:32:49.085677 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.2292 (* 0.0272727 = 0.0335235 loss) | |
| I0525 04:32:49.085691 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.463374 (* 0.0272727 = 0.0126375 loss) | |
| I0525 04:32:49.085705 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.121401 (* 0.0272727 = 0.00331094 loss) | |
| I0525 04:32:49.085719 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.0719074 (* 0.0272727 = 0.00196111 loss) | |
| I0525 04:32:49.085733 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.025424 (* 0.0272727 = 0.000693383 loss) | |
| I0525 04:32:49.085747 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.0189049 (* 0.0272727 = 0.000515587 loss) | |
| I0525 04:32:49.085762 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.0151439 (* 0.0272727 = 0.000413015 loss) | |
| I0525 04:32:49.085799 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.0108729 (* 0.0272727 = 0.000296533 loss) | |
| I0525 04:32:49.085814 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.00773051 (* 0.0272727 = 0.000210832 loss) | |
| I0525 04:32:49.085829 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.0056384 (* 0.0272727 = 0.000153775 loss) | |
| I0525 04:32:49.085842 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.00392531 (* 0.0272727 = 0.000107054 loss) | |
| I0525 04:32:49.085856 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.00281712 (* 0.0272727 = 7.68305e-05 loss) | |
| I0525 04:32:49.085870 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.00259081 (* 0.0272727 = 7.06584e-05 loss) | |
| I0525 04:32:49.085888 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00238352 (* 0.0272727 = 6.50051e-05 loss) | |
| I0525 04:32:49.085902 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00227323 (* 0.0272727 = 6.19972e-05 loss) | |
| I0525 04:32:49.085916 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00221379 (* 0.0272727 = 6.03762e-05 loss) | |
| I0525 04:32:49.085928 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0669266 | |
| I0525 04:32:49.085940 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.123 | |
| I0525 04:32:49.085952 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.105 | |
| I0525 04:32:49.085963 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.083 | |
| I0525 04:32:49.085974 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.167 | |
| I0525 04:32:49.085986 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.341 | |
| I0525 04:32:49.085999 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.481 | |
| I0525 04:32:49.086009 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.736 | |
| I0525 04:32:49.086020 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.918 | |
| I0525 04:32:49.086032 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.98 | |
| I0525 04:32:49.086043 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.993 | |
| I0525 04:32:49.086055 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 1 | |
| I0525 04:32:49.086066 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 1 | |
| I0525 04:32:49.086076 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 1 | |
| I0525 04:32:49.086087 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 1 | |
| I0525 04:32:49.086098 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 1 | |
| I0525 04:32:49.086109 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 1 | |
| I0525 04:32:49.086120 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 1 | |
| I0525 04:32:49.086132 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 1 | |
| I0525 04:32:49.086143 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 1 | |
| I0525 04:32:49.086153 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0525 04:32:49.086164 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0525 04:32:49.086175 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0525 04:32:49.086185 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.766001 | |
| I0525 04:32:49.086197 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.234893 | |
| I0525 04:32:49.086210 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.4892 (* 0.3 = 1.04676 loss) | |
| I0525 04:32:49.086223 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 0.903918 (* 0.3 = 0.271175 loss) | |
| I0525 04:32:49.086237 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 2.84473 (* 0.0272727 = 0.0775836 loss) | |
| I0525 04:32:49.086251 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.0509 (* 0.0272727 = 0.0832063 loss) | |
| I0525 04:32:49.086264 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.14582 (* 0.0272727 = 0.0857952 loss) | |
| I0525 04:32:49.086293 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 2.99103 (* 0.0272727 = 0.0815736 loss) | |
| I0525 04:32:49.086308 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 2.4765 (* 0.0272727 = 0.0675408 loss) | |
| I0525 04:32:49.086323 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.03532 (* 0.0272727 = 0.0555088 loss) | |
| I0525 04:32:49.086335 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.19396 (* 0.0272727 = 0.0325624 loss) | |
| I0525 04:32:49.086349 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.445675 (* 0.0272727 = 0.0121548 loss) | |
| I0525 04:32:49.086364 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.107036 (* 0.0272727 = 0.00291916 loss) | |
| I0525 04:32:49.086376 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.0621045 (* 0.0272727 = 0.00169376 loss) | |
| I0525 04:32:49.086390 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.018584 (* 0.0272727 = 0.000506836 loss) | |
| I0525 04:32:49.086403 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.013575 (* 0.0272727 = 0.000370226 loss) | |
| I0525 04:32:49.086416 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.00972601 (* 0.0272727 = 0.000265255 loss) | |
| I0525 04:32:49.086427 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.00740712 (* 0.0272727 = 0.000202012 loss) | |
| I0525 04:32:49.086436 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.00544423 (* 0.0272727 = 0.000148479 loss) | |
| I0525 04:32:49.086450 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.00405647 (* 0.0272727 = 0.000110631 loss) | |
| I0525 04:32:49.086464 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.00259628 (* 0.0272727 = 7.08076e-05 loss) | |
| I0525 04:32:49.086478 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.00233311 (* 0.0272727 = 6.36302e-05 loss) | |
| I0525 04:32:49.086491 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.00192376 (* 0.0272727 = 5.2466e-05 loss) | |
| I0525 04:32:49.086505 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.0017468 (* 0.0272727 = 4.76401e-05 loss) | |
| I0525 04:32:49.086519 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00161749 (* 0.0272727 = 4.41134e-05 loss) | |
| I0525 04:32:49.086532 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00167208 (* 0.0272727 = 4.56023e-05 loss) | |
| I0525 04:32:49.086544 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0837575 | |
| I0525 04:32:49.086556 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.131 | |
| I0525 04:32:49.086567 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.101 | |
| I0525 04:32:49.086580 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.084 | |
| I0525 04:32:49.086591 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.163 | |
| I0525 04:32:49.086602 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.335 | |
| I0525 04:32:49.086613 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.486 | |
| I0525 04:32:49.086624 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.738 | |
| I0525 04:32:49.086637 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.916 | |
| I0525 04:32:49.086647 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.979 | |
| I0525 04:32:49.086658 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.99 | |
| I0525 04:32:49.086669 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.999 | |
| I0525 04:32:49.086680 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.999 | |
| I0525 04:32:49.086691 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 1 | |
| I0525 04:32:49.086704 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 1 | |
| I0525 04:32:49.086714 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 1 | |
| I0525 04:32:49.086725 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 1 | |
| I0525 04:32:49.086736 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 1 | |
| I0525 04:32:49.086756 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 1 | |
| I0525 04:32:49.086768 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 1 | |
| I0525 04:32:49.086779 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0525 04:32:49.086791 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0525 04:32:49.086802 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0525 04:32:49.086812 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.765546 | |
| I0525 04:32:49.086823 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.266575 | |
| I0525 04:32:49.086838 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.02285 (* 1 = 3.02285 loss) | |
| I0525 04:32:49.086850 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.816482 (* 1 = 0.816482 loss) | |
| I0525 04:32:49.086864 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 2.72822 (* 0.0909091 = 0.24802 loss) | |
| I0525 04:32:49.086877 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 2.96743 (* 0.0909091 = 0.269766 loss) | |
| I0525 04:32:49.086891 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.05545 (* 0.0909091 = 0.277768 loss) | |
| I0525 04:32:49.086905 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 2.90534 (* 0.0909091 = 0.264122 loss) | |
| I0525 04:32:49.086917 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.38122 (* 0.0909091 = 0.216475 loss) | |
| I0525 04:32:49.086933 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 1.93345 (* 0.0909091 = 0.175768 loss) | |
| I0525 04:32:49.086947 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.11479 (* 0.0909091 = 0.101344 loss) | |
| I0525 04:32:49.086961 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.422952 (* 0.0909091 = 0.0384502 loss) | |
| I0525 04:32:49.086974 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.10682 (* 0.0909091 = 0.00971094 loss) | |
| I0525 04:32:49.086988 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.0619808 (* 0.0909091 = 0.00563461 loss) | |
| I0525 04:32:49.087002 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.018001 (* 0.0909091 = 0.00163646 loss) | |
| I0525 04:32:49.087015 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.0124609 (* 0.0909091 = 0.00113281 loss) | |
| I0525 04:32:49.087029 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.00980878 (* 0.0909091 = 0.000891707 loss) | |
| I0525 04:32:49.087043 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.00705698 (* 0.0909091 = 0.000641543 loss) | |
| I0525 04:32:49.087055 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.0053528 (* 0.0909091 = 0.000486618 loss) | |
| I0525 04:32:49.087069 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.00369426 (* 0.0909091 = 0.000335842 loss) | |
| I0525 04:32:49.087082 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.00271653 (* 0.0909091 = 0.000246957 loss) | |
| I0525 04:32:49.087096 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.00195548 (* 0.0909091 = 0.000177771 loss) | |
| I0525 04:32:49.087110 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.00152079 (* 0.0909091 = 0.000138254 loss) | |
| I0525 04:32:49.087123 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00137456 (* 0.0909091 = 0.00012496 loss) | |
| I0525 04:32:49.087137 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.00110642 (* 0.0909091 = 0.000100584 loss) | |
| I0525 04:32:49.087147 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00104875 (* 0.0909091 = 9.53412e-05 loss) | |
| I0525 04:32:49.087159 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 04:32:49.087170 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:32:49.087182 5272 solver.cpp:406] Test net output #149: total_confidence = 0.000255441 | |
| I0525 04:32:49.087193 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000359897 | |
| I0525 04:32:49.087215 5272 solver.cpp:338] Iteration 20000, Testing net (#1) | |
| I0525 04:33:47.102599 5272 solver.cpp:393] Test loss: 9.66398 | |
| I0525 04:33:47.102737 5272 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0714552 | |
| I0525 04:33:47.102758 5272 solver.cpp:406] Test net output #1: loss1/accuracy01 = 0.128 | |
| I0525 04:33:47.102772 5272 solver.cpp:406] Test net output #2: loss1/accuracy02 = 0.115 | |
| I0525 04:33:47.102784 5272 solver.cpp:406] Test net output #3: loss1/accuracy03 = 0.104 | |
| I0525 04:33:47.102797 5272 solver.cpp:406] Test net output #4: loss1/accuracy04 = 0.184 | |
| I0525 04:33:47.102809 5272 solver.cpp:406] Test net output #5: loss1/accuracy05 = 0.34 | |
| I0525 04:33:47.102821 5272 solver.cpp:406] Test net output #6: loss1/accuracy06 = 0.451 | |
| I0525 04:33:47.102833 5272 solver.cpp:406] Test net output #7: loss1/accuracy07 = 0.662 | |
| I0525 04:33:47.102845 5272 solver.cpp:406] Test net output #8: loss1/accuracy08 = 0.826 | |
| I0525 04:33:47.102857 5272 solver.cpp:406] Test net output #9: loss1/accuracy09 = 0.888 | |
| I0525 04:33:47.102869 5272 solver.cpp:406] Test net output #10: loss1/accuracy10 = 0.905 | |
| I0525 04:33:47.102885 5272 solver.cpp:406] Test net output #11: loss1/accuracy11 = 0.926 | |
| I0525 04:33:47.102898 5272 solver.cpp:406] Test net output #12: loss1/accuracy12 = 0.945 | |
| I0525 04:33:47.102910 5272 solver.cpp:406] Test net output #13: loss1/accuracy13 = 0.953 | |
| I0525 04:33:47.102923 5272 solver.cpp:406] Test net output #14: loss1/accuracy14 = 0.964 | |
| I0525 04:33:47.102936 5272 solver.cpp:406] Test net output #15: loss1/accuracy15 = 0.967 | |
| I0525 04:33:47.102947 5272 solver.cpp:406] Test net output #16: loss1/accuracy16 = 0.983 | |
| I0525 04:33:47.102958 5272 solver.cpp:406] Test net output #17: loss1/accuracy17 = 0.993 | |
| I0525 04:33:47.102970 5272 solver.cpp:406] Test net output #18: loss1/accuracy18 = 0.994 | |
| I0525 04:33:47.102982 5272 solver.cpp:406] Test net output #19: loss1/accuracy19 = 0.996 | |
| I0525 04:33:47.102994 5272 solver.cpp:406] Test net output #20: loss1/accuracy20 = 1 | |
| I0525 04:33:47.103005 5272 solver.cpp:406] Test net output #21: loss1/accuracy21 = 1 | |
| I0525 04:33:47.103016 5272 solver.cpp:406] Test net output #22: loss1/accuracy22 = 1 | |
| I0525 04:33:47.103027 5272 solver.cpp:406] Test net output #23: loss1/accuracy_incl_empty = 0.736455 | |
| I0525 04:33:47.103039 5272 solver.cpp:406] Test net output #24: loss1/accuracy_top3 = 0.250323 | |
| I0525 04:33:47.103056 5272 solver.cpp:406] Test net output #25: loss1/cross_entropy_loss = 3.38708 (* 0.3 = 1.01612 loss) | |
| I0525 04:33:47.103070 5272 solver.cpp:406] Test net output #26: loss1/cross_entropy_loss_incl_empty = 0.99432 (* 0.3 = 0.298296 loss) | |
| I0525 04:33:47.103085 5272 solver.cpp:406] Test net output #27: loss1/loss01 = 2.96207 (* 0.0272727 = 0.0807838 loss) | |
| I0525 04:33:47.103098 5272 solver.cpp:406] Test net output #28: loss1/loss02 = 3.05106 (* 0.0272727 = 0.0832107 loss) | |
| I0525 04:33:47.103112 5272 solver.cpp:406] Test net output #29: loss1/loss03 = 3.16663 (* 0.0272727 = 0.0863627 loss) | |
| I0525 04:33:47.103127 5272 solver.cpp:406] Test net output #30: loss1/loss04 = 2.99791 (* 0.0272727 = 0.0817613 loss) | |
| I0525 04:33:47.103139 5272 solver.cpp:406] Test net output #31: loss1/loss05 = 2.53561 (* 0.0272727 = 0.069153 loss) | |
| I0525 04:33:47.103153 5272 solver.cpp:406] Test net output #32: loss1/loss06 = 2.20831 (* 0.0272727 = 0.0602266 loss) | |
| I0525 04:33:47.103168 5272 solver.cpp:406] Test net output #33: loss1/loss07 = 1.49257 (* 0.0272727 = 0.0407064 loss) | |
| I0525 04:33:47.103181 5272 solver.cpp:406] Test net output #34: loss1/loss08 = 0.825123 (* 0.0272727 = 0.0225034 loss) | |
| I0525 04:33:47.103194 5272 solver.cpp:406] Test net output #35: loss1/loss09 = 0.50138 (* 0.0272727 = 0.013674 loss) | |
| I0525 04:33:47.103209 5272 solver.cpp:406] Test net output #36: loss1/loss10 = 0.423965 (* 0.0272727 = 0.0115627 loss) | |
| I0525 04:33:47.103222 5272 solver.cpp:406] Test net output #37: loss1/loss11 = 0.327619 (* 0.0272727 = 0.00893507 loss) | |
| I0525 04:33:47.103236 5272 solver.cpp:406] Test net output #38: loss1/loss12 = 0.256884 (* 0.0272727 = 0.00700594 loss) | |
| I0525 04:33:47.103251 5272 solver.cpp:406] Test net output #39: loss1/loss13 = 0.237937 (* 0.0272727 = 0.0064892 loss) | |
| I0525 04:33:47.103286 5272 solver.cpp:406] Test net output #40: loss1/loss14 = 0.189985 (* 0.0272727 = 0.00518141 loss) | |
| I0525 04:33:47.103301 5272 solver.cpp:406] Test net output #41: loss1/loss15 = 0.175711 (* 0.0272727 = 0.00479213 loss) | |
| I0525 04:33:47.103315 5272 solver.cpp:406] Test net output #42: loss1/loss16 = 0.106645 (* 0.0272727 = 0.00290851 loss) | |
| I0525 04:33:47.103329 5272 solver.cpp:406] Test net output #43: loss1/loss17 = 0.0531984 (* 0.0272727 = 0.00145087 loss) | |
| I0525 04:33:47.103343 5272 solver.cpp:406] Test net output #44: loss1/loss18 = 0.0466086 (* 0.0272727 = 0.00127114 loss) | |
| I0525 04:33:47.103356 5272 solver.cpp:406] Test net output #45: loss1/loss19 = 0.0329337 (* 0.0272727 = 0.000898191 loss) | |
| I0525 04:33:47.103371 5272 solver.cpp:406] Test net output #46: loss1/loss20 = 0.00349135 (* 0.0272727 = 9.52186e-05 loss) | |
| I0525 04:33:47.103385 5272 solver.cpp:406] Test net output #47: loss1/loss21 = 0.00317343 (* 0.0272727 = 8.6548e-05 loss) | |
| I0525 04:33:47.103399 5272 solver.cpp:406] Test net output #48: loss1/loss22 = 0.00299223 (* 0.0272727 = 8.16061e-05 loss) | |
| I0525 04:33:47.103411 5272 solver.cpp:406] Test net output #49: loss2/accuracy = 0.0629042 | |
| I0525 04:33:47.103425 5272 solver.cpp:406] Test net output #50: loss2/accuracy01 = 0.109 | |
| I0525 04:33:47.103436 5272 solver.cpp:406] Test net output #51: loss2/accuracy02 = 0.12 | |
| I0525 04:33:47.103447 5272 solver.cpp:406] Test net output #52: loss2/accuracy03 = 0.086 | |
| I0525 04:33:47.103459 5272 solver.cpp:406] Test net output #53: loss2/accuracy04 = 0.183 | |
| I0525 04:33:47.103471 5272 solver.cpp:406] Test net output #54: loss2/accuracy05 = 0.346 | |
| I0525 04:33:47.103482 5272 solver.cpp:406] Test net output #55: loss2/accuracy06 = 0.448 | |
| I0525 04:33:47.103493 5272 solver.cpp:406] Test net output #56: loss2/accuracy07 = 0.66 | |
| I0525 04:33:47.103505 5272 solver.cpp:406] Test net output #57: loss2/accuracy08 = 0.825 | |
| I0525 04:33:47.103516 5272 solver.cpp:406] Test net output #58: loss2/accuracy09 = 0.889 | |
| I0525 04:33:47.103528 5272 solver.cpp:406] Test net output #59: loss2/accuracy10 = 0.904 | |
| I0525 04:33:47.103539 5272 solver.cpp:406] Test net output #60: loss2/accuracy11 = 0.926 | |
| I0525 04:33:47.103550 5272 solver.cpp:406] Test net output #61: loss2/accuracy12 = 0.945 | |
| I0525 04:33:47.103562 5272 solver.cpp:406] Test net output #62: loss2/accuracy13 = 0.953 | |
| I0525 04:33:47.103574 5272 solver.cpp:406] Test net output #63: loss2/accuracy14 = 0.964 | |
| I0525 04:33:47.103585 5272 solver.cpp:406] Test net output #64: loss2/accuracy15 = 0.967 | |
| I0525 04:33:47.103596 5272 solver.cpp:406] Test net output #65: loss2/accuracy16 = 0.983 | |
| I0525 04:33:47.103607 5272 solver.cpp:406] Test net output #66: loss2/accuracy17 = 0.993 | |
| I0525 04:33:47.103620 5272 solver.cpp:406] Test net output #67: loss2/accuracy18 = 0.994 | |
| I0525 04:33:47.103631 5272 solver.cpp:406] Test net output #68: loss2/accuracy19 = 0.996 | |
| I0525 04:33:47.103641 5272 solver.cpp:406] Test net output #69: loss2/accuracy20 = 1 | |
| I0525 04:33:47.103653 5272 solver.cpp:406] Test net output #70: loss2/accuracy21 = 1 | |
| I0525 04:33:47.103664 5272 solver.cpp:406] Test net output #71: loss2/accuracy22 = 1 | |
| I0525 04:33:47.103675 5272 solver.cpp:406] Test net output #72: loss2/accuracy_incl_empty = 0.73491 | |
| I0525 04:33:47.103687 5272 solver.cpp:406] Test net output #73: loss2/accuracy_top3 = 0.233724 | |
| I0525 04:33:47.103700 5272 solver.cpp:406] Test net output #74: loss2/cross_entropy_loss = 3.48286 (* 0.3 = 1.04486 loss) | |
| I0525 04:33:47.103713 5272 solver.cpp:406] Test net output #75: loss2/cross_entropy_loss_incl_empty = 1.01859 (* 0.3 = 0.305577 loss) | |
| I0525 04:33:47.103727 5272 solver.cpp:406] Test net output #76: loss2/loss01 = 2.9382 (* 0.0272727 = 0.0801327 loss) | |
| I0525 04:33:47.103740 5272 solver.cpp:406] Test net output #77: loss2/loss02 = 3.01608 (* 0.0272727 = 0.0822569 loss) | |
| I0525 04:33:47.103770 5272 solver.cpp:406] Test net output #78: loss2/loss03 = 3.13943 (* 0.0272727 = 0.0856208 loss) | |
| I0525 04:33:47.103785 5272 solver.cpp:406] Test net output #79: loss2/loss04 = 2.97076 (* 0.0272727 = 0.0810207 loss) | |
| I0525 04:33:47.103798 5272 solver.cpp:406] Test net output #80: loss2/loss05 = 2.49861 (* 0.0272727 = 0.0681438 loss) | |
| I0525 04:33:47.103811 5272 solver.cpp:406] Test net output #81: loss2/loss06 = 2.16789 (* 0.0272727 = 0.0591243 loss) | |
| I0525 04:33:47.103824 5272 solver.cpp:406] Test net output #82: loss2/loss07 = 1.46665 (* 0.0272727 = 0.0399995 loss) | |
| I0525 04:33:47.103838 5272 solver.cpp:406] Test net output #83: loss2/loss08 = 0.806178 (* 0.0272727 = 0.0219867 loss) | |
| I0525 04:33:47.103852 5272 solver.cpp:406] Test net output #84: loss2/loss09 = 0.490652 (* 0.0272727 = 0.0133814 loss) | |
| I0525 04:33:47.103865 5272 solver.cpp:406] Test net output #85: loss2/loss10 = 0.412118 (* 0.0272727 = 0.0112396 loss) | |
| I0525 04:33:47.103879 5272 solver.cpp:406] Test net output #86: loss2/loss11 = 0.319816 (* 0.0272727 = 0.00872224 loss) | |
| I0525 04:33:47.103893 5272 solver.cpp:406] Test net output #87: loss2/loss12 = 0.25533 (* 0.0272727 = 0.00696355 loss) | |
| I0525 04:33:47.103907 5272 solver.cpp:406] Test net output #88: loss2/loss13 = 0.23316 (* 0.0272727 = 0.0063589 loss) | |
| I0525 04:33:47.103921 5272 solver.cpp:406] Test net output #89: loss2/loss14 = 0.182048 (* 0.0272727 = 0.00496496 loss) | |
| I0525 04:33:47.103937 5272 solver.cpp:406] Test net output #90: loss2/loss15 = 0.178098 (* 0.0272727 = 0.00485721 loss) | |
| I0525 04:33:47.103952 5272 solver.cpp:406] Test net output #91: loss2/loss16 = 0.103419 (* 0.0272727 = 0.00282052 loss) | |
| I0525 04:33:47.103961 5272 solver.cpp:406] Test net output #92: loss2/loss17 = 0.0530165 (* 0.0272727 = 0.00144591 loss) | |
| I0525 04:33:47.103971 5272 solver.cpp:406] Test net output #93: loss2/loss18 = 0.0476508 (* 0.0272727 = 0.00129957 loss) | |
| I0525 04:33:47.103981 5272 solver.cpp:406] Test net output #94: loss2/loss19 = 0.0333936 (* 0.0272727 = 0.000910735 loss) | |
| I0525 04:33:47.103996 5272 solver.cpp:406] Test net output #95: loss2/loss20 = 0.00289656 (* 0.0272727 = 7.89971e-05 loss) | |
| I0525 04:33:47.104010 5272 solver.cpp:406] Test net output #96: loss2/loss21 = 0.00266811 (* 0.0272727 = 7.27667e-05 loss) | |
| I0525 04:33:47.104023 5272 solver.cpp:406] Test net output #97: loss2/loss22 = 0.00270785 (* 0.0272727 = 7.38505e-05 loss) | |
| I0525 04:33:47.104035 5272 solver.cpp:406] Test net output #98: loss3/accuracy = 0.0863838 | |
| I0525 04:33:47.104048 5272 solver.cpp:406] Test net output #99: loss3/accuracy01 = 0.122 | |
| I0525 04:33:47.104059 5272 solver.cpp:406] Test net output #100: loss3/accuracy02 = 0.107 | |
| I0525 04:33:47.104070 5272 solver.cpp:406] Test net output #101: loss3/accuracy03 = 0.096 | |
| I0525 04:33:47.104082 5272 solver.cpp:406] Test net output #102: loss3/accuracy04 = 0.18 | |
| I0525 04:33:47.104094 5272 solver.cpp:406] Test net output #103: loss3/accuracy05 = 0.342 | |
| I0525 04:33:47.104105 5272 solver.cpp:406] Test net output #104: loss3/accuracy06 = 0.461 | |
| I0525 04:33:47.104116 5272 solver.cpp:406] Test net output #105: loss3/accuracy07 = 0.661 | |
| I0525 04:33:47.104128 5272 solver.cpp:406] Test net output #106: loss3/accuracy08 = 0.827 | |
| I0525 04:33:47.104140 5272 solver.cpp:406] Test net output #107: loss3/accuracy09 = 0.892 | |
| I0525 04:33:47.104151 5272 solver.cpp:406] Test net output #108: loss3/accuracy10 = 0.906 | |
| I0525 04:33:47.104162 5272 solver.cpp:406] Test net output #109: loss3/accuracy11 = 0.928 | |
| I0525 04:33:47.104174 5272 solver.cpp:406] Test net output #110: loss3/accuracy12 = 0.947 | |
| I0525 04:33:47.104185 5272 solver.cpp:406] Test net output #111: loss3/accuracy13 = 0.953 | |
| I0525 04:33:47.104197 5272 solver.cpp:406] Test net output #112: loss3/accuracy14 = 0.964 | |
| I0525 04:33:47.104208 5272 solver.cpp:406] Test net output #113: loss3/accuracy15 = 0.967 | |
| I0525 04:33:47.104219 5272 solver.cpp:406] Test net output #114: loss3/accuracy16 = 0.983 | |
| I0525 04:33:47.104240 5272 solver.cpp:406] Test net output #115: loss3/accuracy17 = 0.993 | |
| I0525 04:33:47.104252 5272 solver.cpp:406] Test net output #116: loss3/accuracy18 = 0.994 | |
| I0525 04:33:47.104264 5272 solver.cpp:406] Test net output #117: loss3/accuracy19 = 0.996 | |
| I0525 04:33:47.104275 5272 solver.cpp:406] Test net output #118: loss3/accuracy20 = 1 | |
| I0525 04:33:47.104286 5272 solver.cpp:406] Test net output #119: loss3/accuracy21 = 1 | |
| I0525 04:33:47.104297 5272 solver.cpp:406] Test net output #120: loss3/accuracy22 = 1 | |
| I0525 04:33:47.104308 5272 solver.cpp:406] Test net output #121: loss3/accuracy_incl_empty = 0.737728 | |
| I0525 04:33:47.104320 5272 solver.cpp:406] Test net output #122: loss3/accuracy_top3 = 0.249018 | |
| I0525 04:33:47.104332 5272 solver.cpp:406] Test net output #123: loss3/cross_entropy_loss = 3.05032 (* 1 = 3.05032 loss) | |
| I0525 04:33:47.104346 5272 solver.cpp:406] Test net output #124: loss3/cross_entropy_loss_incl_empty = 0.931632 (* 1 = 0.931632 loss) | |
| I0525 04:33:47.104359 5272 solver.cpp:406] Test net output #125: loss3/loss01 = 2.83186 (* 0.0909091 = 0.257442 loss) | |
| I0525 04:33:47.104372 5272 solver.cpp:406] Test net output #126: loss3/loss02 = 2.91906 (* 0.0909091 = 0.265369 loss) | |
| I0525 04:33:47.104385 5272 solver.cpp:406] Test net output #127: loss3/loss03 = 3.03042 (* 0.0909091 = 0.275493 loss) | |
| I0525 04:33:47.104399 5272 solver.cpp:406] Test net output #128: loss3/loss04 = 2.87507 (* 0.0909091 = 0.26137 loss) | |
| I0525 04:33:47.104413 5272 solver.cpp:406] Test net output #129: loss3/loss05 = 2.39901 (* 0.0909091 = 0.218091 loss) | |
| I0525 04:33:47.104425 5272 solver.cpp:406] Test net output #130: loss3/loss06 = 2.0483 (* 0.0909091 = 0.18621 loss) | |
| I0525 04:33:47.104439 5272 solver.cpp:406] Test net output #131: loss3/loss07 = 1.36443 (* 0.0909091 = 0.124039 loss) | |
| I0525 04:33:47.104452 5272 solver.cpp:406] Test net output #132: loss3/loss08 = 0.761142 (* 0.0909091 = 0.0691948 loss) | |
| I0525 04:33:47.104465 5272 solver.cpp:406] Test net output #133: loss3/loss09 = 0.451978 (* 0.0909091 = 0.0410889 loss) | |
| I0525 04:33:47.104480 5272 solver.cpp:406] Test net output #134: loss3/loss10 = 0.386661 (* 0.0909091 = 0.035151 loss) | |
| I0525 04:33:47.104493 5272 solver.cpp:406] Test net output #135: loss3/loss11 = 0.28967 (* 0.0909091 = 0.0263336 loss) | |
| I0525 04:33:47.104506 5272 solver.cpp:406] Test net output #136: loss3/loss12 = 0.228419 (* 0.0909091 = 0.0207653 loss) | |
| I0525 04:33:47.104519 5272 solver.cpp:406] Test net output #137: loss3/loss13 = 0.210345 (* 0.0909091 = 0.0191223 loss) | |
| I0525 04:33:47.104533 5272 solver.cpp:406] Test net output #138: loss3/loss14 = 0.16126 (* 0.0909091 = 0.01466 loss) | |
| I0525 04:33:47.104547 5272 solver.cpp:406] Test net output #139: loss3/loss15 = 0.151315 (* 0.0909091 = 0.0137559 loss) | |
| I0525 04:33:47.104560 5272 solver.cpp:406] Test net output #140: loss3/loss16 = 0.0887068 (* 0.0909091 = 0.00806425 loss) | |
| I0525 04:33:47.104573 5272 solver.cpp:406] Test net output #141: loss3/loss17 = 0.042288 (* 0.0909091 = 0.00384436 loss) | |
| I0525 04:33:47.104588 5272 solver.cpp:406] Test net output #142: loss3/loss18 = 0.0391222 (* 0.0909091 = 0.00355657 loss) | |
| I0525 04:33:47.104600 5272 solver.cpp:406] Test net output #143: loss3/loss19 = 0.0276054 (* 0.0909091 = 0.00250958 loss) | |
| I0525 04:33:47.104614 5272 solver.cpp:406] Test net output #144: loss3/loss20 = 0.00243301 (* 0.0909091 = 0.000221183 loss) | |
| I0525 04:33:47.104629 5272 solver.cpp:406] Test net output #145: loss3/loss21 = 0.0016541 (* 0.0909091 = 0.000150373 loss) | |
| I0525 04:33:47.104641 5272 solver.cpp:406] Test net output #146: loss3/loss22 = 0.00130346 (* 0.0909091 = 0.000118496 loss) | |
| I0525 04:33:47.104653 5272 solver.cpp:406] Test net output #147: total_accuracy = 0 | |
| I0525 04:33:47.104665 5272 solver.cpp:406] Test net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:33:47.104676 5272 solver.cpp:406] Test net output #149: total_confidence = 0.000255476 | |
| I0525 04:33:47.104692 5272 solver.cpp:406] Test net output #150: total_confidence_not_rec = 0.000321263 | |
| I0525 04:33:47.463388 5272 solver.cpp:229] Iteration 20000, loss = 9.77133 | |
| I0525 04:33:47.463465 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.131148 | |
| I0525 04:33:47.463486 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 04:33:47.463500 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 04:33:47.463513 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 04:33:47.463526 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 04:33:47.463538 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0525 04:33:47.463551 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 04:33:47.463564 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 04:33:47.463577 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.875 | |
| I0525 04:33:47.463589 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 04:33:47.463603 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 04:33:47.463614 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 04:33:47.463627 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 04:33:47.463640 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 04:33:47.463652 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 04:33:47.463665 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 04:33:47.463676 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 04:33:47.463690 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0525 04:33:47.463701 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0525 04:33:47.463722 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0.875 | |
| I0525 04:33:47.463737 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 0.875 | |
| I0525 04:33:47.463754 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:33:47.463773 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:33:47.463785 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.676136 | |
| I0525 04:33:47.463798 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.245902 | |
| I0525 04:33:47.463814 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.99764 (* 0.3 = 0.899291 loss) | |
| I0525 04:33:47.463829 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.21107 (* 0.3 = 0.363321 loss) | |
| I0525 04:33:47.463843 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.88947 (* 0.0272727 = 0.0788038 loss) | |
| I0525 04:33:47.463858 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.20424 (* 0.0272727 = 0.0873885 loss) | |
| I0525 04:33:47.463872 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.39828 (* 0.0272727 = 0.0926804 loss) | |
| I0525 04:33:47.463886 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.31359 (* 0.0272727 = 0.0903706 loss) | |
| I0525 04:33:47.463901 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.44813 (* 0.0272727 = 0.0940398 loss) | |
| I0525 04:33:47.463914 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.91147 (* 0.0272727 = 0.0794038 loss) | |
| I0525 04:33:47.463928 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.06043 (* 0.0272727 = 0.0561935 loss) | |
| I0525 04:33:47.463943 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.696244 (* 0.0272727 = 0.0189885 loss) | |
| I0525 04:33:47.463956 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.380352 (* 0.0272727 = 0.0103732 loss) | |
| I0525 04:33:47.463970 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.388004 (* 0.0272727 = 0.0105819 loss) | |
| I0525 04:33:47.463984 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.401543 (* 0.0272727 = 0.0109512 loss) | |
| I0525 04:33:47.464028 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.376872 (* 0.0272727 = 0.0102783 loss) | |
| I0525 04:33:47.464045 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.470881 (* 0.0272727 = 0.0128422 loss) | |
| I0525 04:33:47.464058 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.396428 (* 0.0272727 = 0.0108117 loss) | |
| I0525 04:33:47.464072 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.547539 (* 0.0272727 = 0.0149329 loss) | |
| I0525 04:33:47.464098 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.582506 (* 0.0272727 = 0.0158865 loss) | |
| I0525 04:33:47.464126 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.529664 (* 0.0272727 = 0.0144454 loss) | |
| I0525 04:33:47.464143 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.594326 (* 0.0272727 = 0.0162089 loss) | |
| I0525 04:33:47.464157 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.762178 (* 0.0272727 = 0.0207867 loss) | |
| I0525 04:33:47.464170 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.701677 (* 0.0272727 = 0.0191366 loss) | |
| I0525 04:33:47.464187 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.0147884 (* 0.0272727 = 0.00040332 loss) | |
| I0525 04:33:47.464210 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0140148 (* 0.0272727 = 0.000382222 loss) | |
| I0525 04:33:47.464224 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0983607 | |
| I0525 04:33:47.464242 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 04:33:47.464257 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 04:33:47.464270 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 04:33:47.464282 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 04:33:47.464293 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0 | |
| I0525 04:33:47.464306 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.25 | |
| I0525 04:33:47.464318 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.625 | |
| I0525 04:33:47.464330 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.875 | |
| I0525 04:33:47.464342 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 04:33:47.464354 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 04:33:47.464366 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 04:33:47.464378 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 04:33:47.464390 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 04:33:47.464402 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 04:33:47.464414 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 04:33:47.464426 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 04:33:47.464439 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0525 04:33:47.464450 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0525 04:33:47.464462 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0.875 | |
| I0525 04:33:47.464474 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 0.875 | |
| I0525 04:33:47.464486 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:33:47.464498 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:33:47.464509 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.676136 | |
| I0525 04:33:47.464521 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.245902 | |
| I0525 04:33:47.464535 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.11025 (* 0.3 = 0.933076 loss) | |
| I0525 04:33:47.464550 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.18985 (* 0.3 = 0.356954 loss) | |
| I0525 04:33:47.464576 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.89456 (* 0.0272727 = 0.0789426 loss) | |
| I0525 04:33:47.464591 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.85149 (* 0.0272727 = 0.077768 loss) | |
| I0525 04:33:47.464606 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.23511 (* 0.0272727 = 0.0882303 loss) | |
| I0525 04:33:47.464619 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.39444 (* 0.0272727 = 0.0925756 loss) | |
| I0525 04:33:47.464633 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.45539 (* 0.0272727 = 0.0942378 loss) | |
| I0525 04:33:47.464646 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.06186 (* 0.0272727 = 0.0835053 loss) | |
| I0525 04:33:47.464660 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.26448 (* 0.0272727 = 0.0617586 loss) | |
| I0525 04:33:47.464674 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.631294 (* 0.0272727 = 0.0172171 loss) | |
| I0525 04:33:47.464689 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.445314 (* 0.0272727 = 0.0121449 loss) | |
| I0525 04:33:47.464702 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.35271 (* 0.0272727 = 0.00961936 loss) | |
| I0525 04:33:47.464716 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.410395 (* 0.0272727 = 0.0111926 loss) | |
| I0525 04:33:47.464730 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.502393 (* 0.0272727 = 0.0137016 loss) | |
| I0525 04:33:47.464752 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.671797 (* 0.0272727 = 0.0183217 loss) | |
| I0525 04:33:47.464771 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.542856 (* 0.0272727 = 0.0148052 loss) | |
| I0525 04:33:47.464787 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.598881 (* 0.0272727 = 0.0163331 loss) | |
| I0525 04:33:47.464807 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.728908 (* 0.0272727 = 0.0198793 loss) | |
| I0525 04:33:47.464825 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.56225 (* 0.0272727 = 0.0153341 loss) | |
| I0525 04:33:47.464840 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.82687 (* 0.0272727 = 0.022551 loss) | |
| I0525 04:33:47.464854 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.885772 (* 0.0272727 = 0.0241574 loss) | |
| I0525 04:33:47.464869 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 1.02538 (* 0.0272727 = 0.027965 loss) | |
| I0525 04:33:47.464882 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.0215737 (* 0.0272727 = 0.000588374 loss) | |
| I0525 04:33:47.464896 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00956966 (* 0.0272727 = 0.000260991 loss) | |
| I0525 04:33:47.464910 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0819672 | |
| I0525 04:33:47.464922 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 04:33:47.464931 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 04:33:47.464942 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 04:33:47.464956 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 04:33:47.464967 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 04:33:47.464978 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 04:33:47.464990 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 04:33:47.465003 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.875 | |
| I0525 04:33:47.465014 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 04:33:47.465025 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 04:33:47.465037 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 04:33:47.465049 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 04:33:47.465061 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 04:33:47.465085 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 04:33:47.465097 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 04:33:47.465109 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 04:33:47.465142 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0525 04:33:47.465157 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0525 04:33:47.465168 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0.875 | |
| I0525 04:33:47.465181 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 0.875 | |
| I0525 04:33:47.465193 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:33:47.465205 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:33:47.465216 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.664773 | |
| I0525 04:33:47.465229 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.262295 | |
| I0525 04:33:47.465243 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.95942 (* 1 = 2.95942 loss) | |
| I0525 04:33:47.465257 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.17153 (* 1 = 1.17153 loss) | |
| I0525 04:33:47.465271 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.565 (* 0.0909091 = 0.233182 loss) | |
| I0525 04:33:47.465288 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.68415 (* 0.0909091 = 0.244014 loss) | |
| I0525 04:33:47.465311 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.2363 (* 0.0909091 = 0.294209 loss) | |
| I0525 04:33:47.465325 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.19628 (* 0.0909091 = 0.290571 loss) | |
| I0525 04:33:47.465348 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 3.04095 (* 0.0909091 = 0.27645 loss) | |
| I0525 04:33:47.465363 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.40104 (* 0.0909091 = 0.218276 loss) | |
| I0525 04:33:47.465376 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.82544 (* 0.0909091 = 0.165949 loss) | |
| I0525 04:33:47.465390 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.639629 (* 0.0909091 = 0.0581481 loss) | |
| I0525 04:33:47.465404 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.329263 (* 0.0909091 = 0.029933 loss) | |
| I0525 04:33:47.465418 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.293536 (* 0.0909091 = 0.026685 loss) | |
| I0525 04:33:47.465432 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.434204 (* 0.0909091 = 0.039473 loss) | |
| I0525 04:33:47.465446 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.317428 (* 0.0909091 = 0.0288571 loss) | |
| I0525 04:33:47.465461 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.57442 (* 0.0909091 = 0.05222 loss) | |
| I0525 04:33:47.465476 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.665187 (* 0.0909091 = 0.0604715 loss) | |
| I0525 04:33:47.465489 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.674674 (* 0.0909091 = 0.061334 loss) | |
| I0525 04:33:47.465503 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.676936 (* 0.0909091 = 0.0615396 loss) | |
| I0525 04:33:47.465517 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.726908 (* 0.0909091 = 0.0660826 loss) | |
| I0525 04:33:47.465530 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.680123 (* 0.0909091 = 0.0618294 loss) | |
| I0525 04:33:47.465544 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.747721 (* 0.0909091 = 0.0679747 loss) | |
| I0525 04:33:47.465559 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.903039 (* 0.0909091 = 0.0820944 loss) | |
| I0525 04:33:47.465574 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000866617 (* 0.0909091 = 7.87834e-05 loss) | |
| I0525 04:33:47.465587 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000338962 (* 0.0909091 = 3.08148e-05 loss) | |
| I0525 04:33:47.465600 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:33:47.465625 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:33:47.465637 5272 solver.cpp:245] Train net output #149: total_confidence = 4.79402e-07 | |
| I0525 04:33:47.465649 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 4.62061e-06 | |
| I0525 04:33:47.465662 5272 sgd_solver.cpp:106] Iteration 20000, lr = 0.001 | |
| I0525 04:35:17.181689 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.4282 > 30) by scale factor 0.985927 | |
| I0525 04:36:00.312917 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 46.3671 > 30) by scale factor 0.647011 | |
| I0525 04:38:00.449288 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.9566 > 30) by scale factor 0.969099 | |
| I0525 04:39:10.536027 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 44.3731 > 30) by scale factor 0.676085 | |
| I0525 04:39:16.710968 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 36.7165 > 30) by scale factor 0.817072 | |
| I0525 04:40:12.535034 5272 solver.cpp:229] Iteration 20500, loss = 9.7067 | |
| I0525 04:40:12.535162 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.116667 | |
| I0525 04:40:12.535184 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 04:40:12.535197 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 04:40:12.535210 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 04:40:12.535224 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.375 | |
| I0525 04:40:12.535238 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 04:40:12.535250 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.5 | |
| I0525 04:40:12.535262 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 04:40:12.535275 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 04:40:12.535287 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.75 | |
| I0525 04:40:12.535300 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.75 | |
| I0525 04:40:12.535313 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.75 | |
| I0525 04:40:12.535326 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.75 | |
| I0525 04:40:12.535339 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 04:40:12.535351 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 04:40:12.535363 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.875 | |
| I0525 04:40:12.535375 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 04:40:12.535387 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0525 04:40:12.535399 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 0.875 | |
| I0525 04:40:12.535411 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 0.875 | |
| I0525 04:40:12.535423 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:40:12.535435 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:40:12.535447 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:40:12.535459 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.670455 | |
| I0525 04:40:12.535471 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.2 | |
| I0525 04:40:12.535488 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.01785 (* 0.3 = 0.905354 loss) | |
| I0525 04:40:12.535502 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.22847 (* 0.3 = 0.368542 loss) | |
| I0525 04:40:12.535517 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.24312 (* 0.0272727 = 0.0884487 loss) | |
| I0525 04:40:12.535531 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 2.93556 (* 0.0272727 = 0.0800607 loss) | |
| I0525 04:40:12.535545 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.25044 (* 0.0272727 = 0.0886485 loss) | |
| I0525 04:40:12.535559 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.9032 (* 0.0272727 = 0.0791782 loss) | |
| I0525 04:40:12.535573 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.2301 (* 0.0272727 = 0.0608209 loss) | |
| I0525 04:40:12.535588 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.86725 (* 0.0272727 = 0.0509249 loss) | |
| I0525 04:40:12.535600 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.32501 (* 0.0272727 = 0.0634093 loss) | |
| I0525 04:40:12.535614 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.778871 (* 0.0272727 = 0.0212419 loss) | |
| I0525 04:40:12.535629 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.890573 (* 0.0272727 = 0.0242884 loss) | |
| I0525 04:40:12.535642 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.828401 (* 0.0272727 = 0.0225928 loss) | |
| I0525 04:40:12.535656 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 1.38788 (* 0.0272727 = 0.0378514 loss) | |
| I0525 04:40:12.535670 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 1.0536 (* 0.0272727 = 0.0287346 loss) | |
| I0525 04:40:12.535703 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.640633 (* 0.0272727 = 0.0174718 loss) | |
| I0525 04:40:12.535719 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.678007 (* 0.0272727 = 0.0184911 loss) | |
| I0525 04:40:12.535733 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 1.02732 (* 0.0272727 = 0.0280178 loss) | |
| I0525 04:40:12.535748 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.865885 (* 0.0272727 = 0.023615 loss) | |
| I0525 04:40:12.535761 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 1.30067 (* 0.0272727 = 0.0354729 loss) | |
| I0525 04:40:12.535775 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 1.38763 (* 0.0272727 = 0.0378443 loss) | |
| I0525 04:40:12.535789 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 1.21311 (* 0.0272727 = 0.0330849 loss) | |
| I0525 04:40:12.535804 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00473243 (* 0.0272727 = 0.000129066 loss) | |
| I0525 04:40:12.535817 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00159194 (* 0.0272727 = 4.34165e-05 loss) | |
| I0525 04:40:12.535831 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000666336 (* 0.0272727 = 1.81728e-05 loss) | |
| I0525 04:40:12.535845 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0666667 | |
| I0525 04:40:12.535856 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 04:40:12.535868 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 04:40:12.535883 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0525 04:40:12.535897 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 04:40:12.535908 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 04:40:12.535920 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 04:40:12.535933 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 04:40:12.535944 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 04:40:12.535956 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.75 | |
| I0525 04:40:12.535969 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.75 | |
| I0525 04:40:12.535980 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.75 | |
| I0525 04:40:12.535992 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.75 | |
| I0525 04:40:12.536005 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 04:40:12.536016 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 04:40:12.536028 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.875 | |
| I0525 04:40:12.536041 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 04:40:12.536052 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0525 04:40:12.536064 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 0.875 | |
| I0525 04:40:12.536077 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 0.875 | |
| I0525 04:40:12.536087 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:40:12.536099 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:40:12.536110 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:40:12.536123 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.664773 | |
| I0525 04:40:12.536134 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.25 | |
| I0525 04:40:12.536147 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.18994 (* 0.3 = 0.956981 loss) | |
| I0525 04:40:12.536161 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.22881 (* 0.3 = 0.368643 loss) | |
| I0525 04:40:12.536177 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 2.6376 (* 0.0272727 = 0.0719345 loss) | |
| I0525 04:40:12.536192 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 2.97868 (* 0.0272727 = 0.0812366 loss) | |
| I0525 04:40:12.536216 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 2.9217 (* 0.0272727 = 0.0796827 loss) | |
| I0525 04:40:12.536231 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.03985 (* 0.0272727 = 0.0829051 loss) | |
| I0525 04:40:12.536245 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.0999 (* 0.0272727 = 0.0572701 loss) | |
| I0525 04:40:12.536259 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.28378 (* 0.0272727 = 0.0622848 loss) | |
| I0525 04:40:12.536273 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.1726 (* 0.0272727 = 0.0592528 loss) | |
| I0525 04:40:12.536288 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.03798 (* 0.0272727 = 0.0283087 loss) | |
| I0525 04:40:12.536300 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.01961 (* 0.0272727 = 0.0278076 loss) | |
| I0525 04:40:12.536314 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.95397 (* 0.0272727 = 0.0260174 loss) | |
| I0525 04:40:12.536329 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 1.09445 (* 0.0272727 = 0.0298486 loss) | |
| I0525 04:40:12.536342 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.880937 (* 0.0272727 = 0.0240256 loss) | |
| I0525 04:40:12.536356 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.564397 (* 0.0272727 = 0.0153926 loss) | |
| I0525 04:40:12.536370 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.61983 (* 0.0272727 = 0.0169044 loss) | |
| I0525 04:40:12.536384 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.783707 (* 0.0272727 = 0.0213738 loss) | |
| I0525 04:40:12.536398 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.518647 (* 0.0272727 = 0.0141449 loss) | |
| I0525 04:40:12.536412 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.987281 (* 0.0272727 = 0.0269258 loss) | |
| I0525 04:40:12.536425 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 1.05975 (* 0.0272727 = 0.0289024 loss) | |
| I0525 04:40:12.536439 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.96476 (* 0.0272727 = 0.0263116 loss) | |
| I0525 04:40:12.536453 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00230762 (* 0.0272727 = 6.2935e-05 loss) | |
| I0525 04:40:12.536468 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00268712 (* 0.0272727 = 7.32851e-05 loss) | |
| I0525 04:40:12.536481 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00171611 (* 0.0272727 = 4.68031e-05 loss) | |
| I0525 04:40:12.536494 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0833333 | |
| I0525 04:40:12.536505 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 04:40:12.536517 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0 | |
| I0525 04:40:12.536528 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 04:40:12.536540 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 04:40:12.536552 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 04:40:12.536564 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.5 | |
| I0525 04:40:12.536576 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.625 | |
| I0525 04:40:12.536588 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 04:40:12.536600 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.75 | |
| I0525 04:40:12.536612 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.75 | |
| I0525 04:40:12.536624 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.75 | |
| I0525 04:40:12.536635 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.75 | |
| I0525 04:40:12.536648 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 04:40:12.536659 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 04:40:12.536671 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.875 | |
| I0525 04:40:12.536682 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 04:40:12.536705 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0525 04:40:12.536717 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 0.875 | |
| I0525 04:40:12.536730 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 0.875 | |
| I0525 04:40:12.536741 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:40:12.536753 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:40:12.536764 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:40:12.536777 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.676136 | |
| I0525 04:40:12.536788 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.3 | |
| I0525 04:40:12.536803 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.01044 (* 1 = 3.01044 loss) | |
| I0525 04:40:12.536816 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.18061 (* 1 = 1.18061 loss) | |
| I0525 04:40:12.536829 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.47921 (* 0.0909091 = 0.225383 loss) | |
| I0525 04:40:12.536844 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.00676 (* 0.0909091 = 0.273342 loss) | |
| I0525 04:40:12.536857 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.0504 (* 0.0909091 = 0.277309 loss) | |
| I0525 04:40:12.536871 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 2.95546 (* 0.0909091 = 0.268678 loss) | |
| I0525 04:40:12.536885 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 1.98036 (* 0.0909091 = 0.180033 loss) | |
| I0525 04:40:12.536900 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.02785 (* 0.0909091 = 0.18435 loss) | |
| I0525 04:40:12.536912 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 1.77415 (* 0.0909091 = 0.161287 loss) | |
| I0525 04:40:12.536929 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.792669 (* 0.0909091 = 0.0720608 loss) | |
| I0525 04:40:12.536944 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.876183 (* 0.0909091 = 0.079653 loss) | |
| I0525 04:40:12.536958 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.846718 (* 0.0909091 = 0.0769743 loss) | |
| I0525 04:40:12.536972 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.700301 (* 0.0909091 = 0.0636638 loss) | |
| I0525 04:40:12.536985 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.68302 (* 0.0909091 = 0.0620927 loss) | |
| I0525 04:40:12.537001 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.551978 (* 0.0909091 = 0.0501798 loss) | |
| I0525 04:40:12.537011 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.565866 (* 0.0909091 = 0.0514423 loss) | |
| I0525 04:40:12.537025 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.718369 (* 0.0909091 = 0.0653063 loss) | |
| I0525 04:40:12.537039 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.7799 (* 0.0909091 = 0.0709 loss) | |
| I0525 04:40:12.537053 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.617818 (* 0.0909091 = 0.0561653 loss) | |
| I0525 04:40:12.537067 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.96885 (* 0.0909091 = 0.0880772 loss) | |
| I0525 04:40:12.537081 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 1.08615 (* 0.0909091 = 0.098741 loss) | |
| I0525 04:40:12.537096 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00462998 (* 0.0909091 = 0.000420907 loss) | |
| I0525 04:40:12.537109 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00122286 (* 0.0909091 = 0.000111169 loss) | |
| I0525 04:40:12.537142 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000808319 (* 0.0909091 = 7.34836e-05 loss) | |
| I0525 04:40:12.537155 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:40:12.537168 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:40:12.537179 5272 solver.cpp:245] Train net output #149: total_confidence = 0.000137818 | |
| I0525 04:40:12.537202 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.00027517 | |
| I0525 04:40:12.537217 5272 sgd_solver.cpp:106] Iteration 20500, lr = 0.001 | |
| I0525 04:43:26.946056 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.5656 > 30) by scale factor 0.981495 | |
| I0525 04:46:37.703830 5272 solver.cpp:229] Iteration 21000, loss = 9.69837 | |
| I0525 04:46:37.703990 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0681818 | |
| I0525 04:46:37.704013 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 04:46:37.704028 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.125 | |
| I0525 04:46:37.704041 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0.125 | |
| I0525 04:46:37.704053 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 04:46:37.704066 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.625 | |
| I0525 04:46:37.704078 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.375 | |
| I0525 04:46:37.704090 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 04:46:37.704103 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 04:46:37.704116 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 04:46:37.704128 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 04:46:37.704141 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 04:46:37.704154 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 04:46:37.704166 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 04:46:37.704177 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 04:46:37.704190 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 04:46:37.704201 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 04:46:37.704213 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 04:46:37.704226 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:46:37.704237 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:46:37.704249 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:46:37.704260 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:46:37.704272 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:46:37.704284 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.75 | |
| I0525 04:46:37.704298 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.227273 | |
| I0525 04:46:37.704313 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.21026 (* 0.3 = 0.963079 loss) | |
| I0525 04:46:37.704329 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.927103 (* 0.3 = 0.278131 loss) | |
| I0525 04:46:37.704344 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.34841 (* 0.0272727 = 0.0913202 loss) | |
| I0525 04:46:37.704357 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.36648 (* 0.0272727 = 0.091813 loss) | |
| I0525 04:46:37.704371 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.08953 (* 0.0272727 = 0.08426 loss) | |
| I0525 04:46:37.704385 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.81931 (* 0.0272727 = 0.104163 loss) | |
| I0525 04:46:37.704399 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.25283 (* 0.0272727 = 0.0614409 loss) | |
| I0525 04:46:37.704414 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.50128 (* 0.0272727 = 0.0682166 loss) | |
| I0525 04:46:37.704428 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 0.806326 (* 0.0272727 = 0.0219907 loss) | |
| I0525 04:46:37.704443 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0654782 (* 0.0272727 = 0.00178577 loss) | |
| I0525 04:46:37.704458 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0437953 (* 0.0272727 = 0.00119442 loss) | |
| I0525 04:46:37.704473 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0237511 (* 0.0272727 = 0.000647758 loss) | |
| I0525 04:46:37.704488 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.0295489 (* 0.0272727 = 0.000805878 loss) | |
| I0525 04:46:37.704501 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.0268581 (* 0.0272727 = 0.000732494 loss) | |
| I0525 04:46:37.704516 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.0174152 (* 0.0272727 = 0.000474959 loss) | |
| I0525 04:46:37.704552 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.0240333 (* 0.0272727 = 0.000655453 loss) | |
| I0525 04:46:37.704567 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0123189 (* 0.0272727 = 0.000335969 loss) | |
| I0525 04:46:37.704581 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.0077471 (* 0.0272727 = 0.000211285 loss) | |
| I0525 04:46:37.704596 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.0160953 (* 0.0272727 = 0.000438963 loss) | |
| I0525 04:46:37.704609 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00849851 (* 0.0272727 = 0.000231778 loss) | |
| I0525 04:46:37.704624 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0111119 (* 0.0272727 = 0.000303051 loss) | |
| I0525 04:46:37.704638 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00658314 (* 0.0272727 = 0.00017954 loss) | |
| I0525 04:46:37.704653 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00659746 (* 0.0272727 = 0.000179931 loss) | |
| I0525 04:46:37.704666 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00511409 (* 0.0272727 = 0.000139475 loss) | |
| I0525 04:46:37.704679 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0681818 | |
| I0525 04:46:37.704691 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 04:46:37.704704 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0.125 | |
| I0525 04:46:37.704715 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.25 | |
| I0525 04:46:37.704727 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 04:46:37.704740 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 04:46:37.704751 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 04:46:37.704763 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 04:46:37.704776 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 04:46:37.704787 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 04:46:37.704798 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 04:46:37.704810 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 04:46:37.704821 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 04:46:37.704833 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 04:46:37.704844 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 04:46:37.704856 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 04:46:37.704869 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 04:46:37.704884 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 04:46:37.704895 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:46:37.704906 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:46:37.704918 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:46:37.704929 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:46:37.704941 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:46:37.704952 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.761364 | |
| I0525 04:46:37.704964 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.204545 | |
| I0525 04:46:37.704979 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.19877 (* 0.3 = 0.959631 loss) | |
| I0525 04:46:37.704989 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.874753 (* 0.3 = 0.262426 loss) | |
| I0525 04:46:37.705008 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.04544 (* 0.0272727 = 0.0830575 loss) | |
| I0525 04:46:37.705021 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.176 (* 0.0272727 = 0.0866181 loss) | |
| I0525 04:46:37.705046 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.04488 (* 0.0272727 = 0.0830422 loss) | |
| I0525 04:46:37.705061 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.78814 (* 0.0272727 = 0.103313 loss) | |
| I0525 04:46:37.705075 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.98152 (* 0.0272727 = 0.0813142 loss) | |
| I0525 04:46:37.705090 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.46513 (* 0.0272727 = 0.0672308 loss) | |
| I0525 04:46:37.705102 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 0.792969 (* 0.0272727 = 0.0216264 loss) | |
| I0525 04:46:37.705129 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.087227 (* 0.0272727 = 0.00237892 loss) | |
| I0525 04:46:37.705147 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.0256489 (* 0.0272727 = 0.000699516 loss) | |
| I0525 04:46:37.705162 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.0297102 (* 0.0272727 = 0.000810277 loss) | |
| I0525 04:46:37.705175 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.0193213 (* 0.0272727 = 0.000526944 loss) | |
| I0525 04:46:37.705189 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.00876074 (* 0.0272727 = 0.000238929 loss) | |
| I0525 04:46:37.705204 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00725171 (* 0.0272727 = 0.000197774 loss) | |
| I0525 04:46:37.705217 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00767833 (* 0.0272727 = 0.000209409 loss) | |
| I0525 04:46:37.705231 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0109744 (* 0.0272727 = 0.000299303 loss) | |
| I0525 04:46:37.705245 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.00755705 (* 0.0272727 = 0.000206101 loss) | |
| I0525 04:46:37.705260 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00389646 (* 0.0272727 = 0.000106267 loss) | |
| I0525 04:46:37.705273 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.00747994 (* 0.0272727 = 0.000203998 loss) | |
| I0525 04:46:37.705287 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.00434774 (* 0.0272727 = 0.000118575 loss) | |
| I0525 04:46:37.705302 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00410322 (* 0.0272727 = 0.000111906 loss) | |
| I0525 04:46:37.705315 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00305338 (* 0.0272727 = 8.32739e-05 loss) | |
| I0525 04:46:37.705329 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00384496 (* 0.0272727 = 0.000104863 loss) | |
| I0525 04:46:37.705341 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0454545 | |
| I0525 04:46:37.705353 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 04:46:37.705365 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0525 04:46:37.705377 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 04:46:37.705389 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 04:46:37.705401 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.25 | |
| I0525 04:46:37.705413 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 04:46:37.705425 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 04:46:37.705436 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 04:46:37.705448 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 04:46:37.705459 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 04:46:37.705471 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 04:46:37.705483 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 04:46:37.705494 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 04:46:37.705507 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 04:46:37.705518 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 04:46:37.705529 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 04:46:37.705551 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 04:46:37.705564 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:46:37.705576 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:46:37.705588 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:46:37.705600 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:46:37.705611 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:46:37.705623 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.761364 | |
| I0525 04:46:37.705636 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.204545 | |
| I0525 04:46:37.705649 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.06587 (* 1 = 3.06587 loss) | |
| I0525 04:46:37.705663 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.810092 (* 1 = 0.810092 loss) | |
| I0525 04:46:37.705677 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.91463 (* 0.0909091 = 0.264967 loss) | |
| I0525 04:46:37.705692 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 2.99078 (* 0.0909091 = 0.271889 loss) | |
| I0525 04:46:37.705705 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.91583 (* 0.0909091 = 0.265075 loss) | |
| I0525 04:46:37.705719 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.50675 (* 0.0909091 = 0.318795 loss) | |
| I0525 04:46:37.705734 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.50299 (* 0.0909091 = 0.227544 loss) | |
| I0525 04:46:37.705744 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.22914 (* 0.0909091 = 0.202649 loss) | |
| I0525 04:46:37.705757 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.589085 (* 0.0909091 = 0.0535532 loss) | |
| I0525 04:46:37.705771 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0697591 (* 0.0909091 = 0.00634174 loss) | |
| I0525 04:46:37.705785 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.0314684 (* 0.0909091 = 0.00286076 loss) | |
| I0525 04:46:37.705799 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.0136882 (* 0.0909091 = 0.00124438 loss) | |
| I0525 04:46:37.705813 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00937546 (* 0.0909091 = 0.000852315 loss) | |
| I0525 04:46:37.705827 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.00902094 (* 0.0909091 = 0.000820086 loss) | |
| I0525 04:46:37.705842 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.00440815 (* 0.0909091 = 0.000400741 loss) | |
| I0525 04:46:37.705855 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.00596475 (* 0.0909091 = 0.00054225 loss) | |
| I0525 04:46:37.705869 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.00431966 (* 0.0909091 = 0.000392697 loss) | |
| I0525 04:46:37.705883 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.00409816 (* 0.0909091 = 0.00037256 loss) | |
| I0525 04:46:37.705896 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.00219154 (* 0.0909091 = 0.000199231 loss) | |
| I0525 04:46:37.705910 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.00198977 (* 0.0909091 = 0.000180888 loss) | |
| I0525 04:46:37.705924 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00170016 (* 0.0909091 = 0.00015456 loss) | |
| I0525 04:46:37.705941 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00145705 (* 0.0909091 = 0.000132459 loss) | |
| I0525 04:46:37.705955 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000985738 (* 0.0909091 = 8.96125e-05 loss) | |
| I0525 04:46:37.705970 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.00116641 (* 0.0909091 = 0.000106037 loss) | |
| I0525 04:46:37.705981 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:46:37.705993 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:46:37.706004 5272 solver.cpp:245] Train net output #149: total_confidence = 0.000343258 | |
| I0525 04:46:37.706027 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000922688 | |
| I0525 04:46:37.706040 5272 sgd_solver.cpp:106] Iteration 21000, lr = 0.001 | |
| I0525 04:49:05.994966 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 42.3085 > 30) by scale factor 0.709078 | |
| I0525 04:49:28.350611 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 35.6114 > 30) by scale factor 0.842428 | |
| I0525 04:50:59.216051 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 41.9117 > 30) by scale factor 0.71579 | |
| I0525 04:51:45.416169 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 43.021 > 30) by scale factor 0.697334 | |
| I0525 04:53:02.799557 5272 solver.cpp:229] Iteration 21500, loss = 9.68776 | |
| I0525 04:53:02.799707 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0945946 | |
| I0525 04:53:02.799728 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.25 | |
| I0525 04:53:02.799742 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 04:53:02.799756 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 04:53:02.799767 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 04:53:02.799779 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0 | |
| I0525 04:53:02.799792 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.25 | |
| I0525 04:53:02.799804 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.25 | |
| I0525 04:53:02.799818 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.375 | |
| I0525 04:53:02.799829 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.625 | |
| I0525 04:53:02.799842 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.625 | |
| I0525 04:53:02.799855 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.75 | |
| I0525 04:53:02.799868 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.75 | |
| I0525 04:53:02.799885 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.75 | |
| I0525 04:53:02.799896 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.75 | |
| I0525 04:53:02.799909 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 0.75 | |
| I0525 04:53:02.799921 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 0.875 | |
| I0525 04:53:02.799933 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 0.875 | |
| I0525 04:53:02.799947 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:53:02.799957 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:53:02.799969 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:53:02.799981 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:53:02.799993 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:53:02.800005 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.590909 | |
| I0525 04:53:02.800017 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.189189 | |
| I0525 04:53:02.800034 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.05676 (* 0.3 = 0.917029 loss) | |
| I0525 04:53:02.800048 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.53411 (* 0.3 = 0.460232 loss) | |
| I0525 04:53:02.800062 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.50438 (* 0.0272727 = 0.095574 loss) | |
| I0525 04:53:02.800076 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.68488 (* 0.0272727 = 0.100497 loss) | |
| I0525 04:53:02.800091 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.22525 (* 0.0272727 = 0.0879613 loss) | |
| I0525 04:53:02.800106 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.83075 (* 0.0272727 = 0.0772022 loss) | |
| I0525 04:53:02.800119 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.94681 (* 0.0272727 = 0.0803675 loss) | |
| I0525 04:53:02.800133 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 2.86366 (* 0.0272727 = 0.0780997 loss) | |
| I0525 04:53:02.800148 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.57267 (* 0.0272727 = 0.0701636 loss) | |
| I0525 04:53:02.800163 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 2.32217 (* 0.0272727 = 0.0633319 loss) | |
| I0525 04:53:02.800176 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 1.44343 (* 0.0272727 = 0.0393663 loss) | |
| I0525 04:53:02.800190 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 1.71323 (* 0.0272727 = 0.0467244 loss) | |
| I0525 04:53:02.800204 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.882942 (* 0.0272727 = 0.0240802 loss) | |
| I0525 04:53:02.800218 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 1.19151 (* 0.0272727 = 0.0324959 loss) | |
| I0525 04:53:02.800256 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 1.06502 (* 0.0272727 = 0.029046 loss) | |
| I0525 04:53:02.800271 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 1.13025 (* 0.0272727 = 0.0308249 loss) | |
| I0525 04:53:02.800284 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 1.39899 (* 0.0272727 = 0.0381543 loss) | |
| I0525 04:53:02.800297 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 1.03353 (* 0.0272727 = 0.0281872 loss) | |
| I0525 04:53:02.800312 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 1.29911 (* 0.0272727 = 0.0354304 loss) | |
| I0525 04:53:02.800325 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0533758 (* 0.0272727 = 0.0014557 loss) | |
| I0525 04:53:02.800340 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00925695 (* 0.0272727 = 0.000252462 loss) | |
| I0525 04:53:02.800354 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.0232706 (* 0.0272727 = 0.000634653 loss) | |
| I0525 04:53:02.800369 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.021859 (* 0.0272727 = 0.000596154 loss) | |
| I0525 04:53:02.800384 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.0263539 (* 0.0272727 = 0.000718742 loss) | |
| I0525 04:53:02.800395 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.162162 | |
| I0525 04:53:02.800407 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0 | |
| I0525 04:53:02.800420 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 04:53:02.800431 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 04:53:02.800443 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 04:53:02.800454 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.25 | |
| I0525 04:53:02.800467 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.375 | |
| I0525 04:53:02.800478 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.25 | |
| I0525 04:53:02.800490 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.5 | |
| I0525 04:53:02.800503 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.625 | |
| I0525 04:53:02.800514 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.625 | |
| I0525 04:53:02.800526 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.75 | |
| I0525 04:53:02.800539 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.75 | |
| I0525 04:53:02.800550 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.75 | |
| I0525 04:53:02.800562 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.75 | |
| I0525 04:53:02.800575 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 0.75 | |
| I0525 04:53:02.800586 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 0.875 | |
| I0525 04:53:02.800598 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 0.875 | |
| I0525 04:53:02.800611 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:53:02.800622 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:53:02.800633 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:53:02.800645 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:53:02.800657 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:53:02.800669 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.625 | |
| I0525 04:53:02.800680 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.22973 | |
| I0525 04:53:02.800695 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.01866 (* 0.3 = 0.905598 loss) | |
| I0525 04:53:02.800709 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.46509 (* 0.3 = 0.439528 loss) | |
| I0525 04:53:02.800726 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.59095 (* 0.0272727 = 0.0979351 loss) | |
| I0525 04:53:02.800740 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.65614 (* 0.0272727 = 0.0997128 loss) | |
| I0525 04:53:02.800766 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.19446 (* 0.0272727 = 0.0871216 loss) | |
| I0525 04:53:02.800781 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.58929 (* 0.0272727 = 0.0978896 loss) | |
| I0525 04:53:02.800796 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.5329 (* 0.0272727 = 0.0690791 loss) | |
| I0525 04:53:02.800808 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 2.96185 (* 0.0272727 = 0.0807777 loss) | |
| I0525 04:53:02.800822 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.86447 (* 0.0272727 = 0.0781218 loss) | |
| I0525 04:53:02.800837 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 2.65774 (* 0.0272727 = 0.0724838 loss) | |
| I0525 04:53:02.800849 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 1.44962 (* 0.0272727 = 0.039535 loss) | |
| I0525 04:53:02.800863 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 1.74077 (* 0.0272727 = 0.0474756 loss) | |
| I0525 04:53:02.800878 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.71616 (* 0.0272727 = 0.0195316 loss) | |
| I0525 04:53:02.800891 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.992351 (* 0.0272727 = 0.0270641 loss) | |
| I0525 04:53:02.800905 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 1.1657 (* 0.0272727 = 0.0317919 loss) | |
| I0525 04:53:02.800920 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.94821 (* 0.0272727 = 0.0258603 loss) | |
| I0525 04:53:02.800936 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 1.50791 (* 0.0272727 = 0.0411249 loss) | |
| I0525 04:53:02.800951 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.936513 (* 0.0272727 = 0.0255413 loss) | |
| I0525 04:53:02.800964 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 1.13222 (* 0.0272727 = 0.0308787 loss) | |
| I0525 04:53:02.800978 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0385646 (* 0.0272727 = 0.00105176 loss) | |
| I0525 04:53:02.800992 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0227015 (* 0.0272727 = 0.000619133 loss) | |
| I0525 04:53:02.801007 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.014601 (* 0.0272727 = 0.00039821 loss) | |
| I0525 04:53:02.801020 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.020092 (* 0.0272727 = 0.000547965 loss) | |
| I0525 04:53:02.801034 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.0109967 (* 0.0272727 = 0.00029991 loss) | |
| I0525 04:53:02.801048 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.108108 | |
| I0525 04:53:02.801059 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0 | |
| I0525 04:53:02.801071 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.375 | |
| I0525 04:53:02.801084 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 04:53:02.801095 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.125 | |
| I0525 04:53:02.801107 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 04:53:02.801131 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.375 | |
| I0525 04:53:02.801146 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.25 | |
| I0525 04:53:02.801158 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.5 | |
| I0525 04:53:02.801170 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.625 | |
| I0525 04:53:02.801182 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.625 | |
| I0525 04:53:02.801194 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.75 | |
| I0525 04:53:02.801206 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 04:53:02.801218 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.75 | |
| I0525 04:53:02.801229 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.75 | |
| I0525 04:53:02.801241 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 0.75 | |
| I0525 04:53:02.801254 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 0.875 | |
| I0525 04:53:02.801277 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 0.875 | |
| I0525 04:53:02.801290 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:53:02.801302 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:53:02.801314 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:53:02.801326 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:53:02.801337 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:53:02.801349 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.602273 | |
| I0525 04:53:02.801362 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.256757 | |
| I0525 04:53:02.801378 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 2.93915 (* 1 = 2.93915 loss) | |
| I0525 04:53:02.801391 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.41913 (* 1 = 1.41913 loss) | |
| I0525 04:53:02.801405 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.08415 (* 0.0909091 = 0.280377 loss) | |
| I0525 04:53:02.801419 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.0482 (* 0.0909091 = 0.277109 loss) | |
| I0525 04:53:02.801434 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 2.9293 (* 0.0909091 = 0.2663 loss) | |
| I0525 04:53:02.801447 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.38845 (* 0.0909091 = 0.308041 loss) | |
| I0525 04:53:02.801461 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.76032 (* 0.0909091 = 0.250938 loss) | |
| I0525 04:53:02.801476 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 2.51969 (* 0.0909091 = 0.229063 loss) | |
| I0525 04:53:02.801488 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.58664 (* 0.0909091 = 0.235149 loss) | |
| I0525 04:53:02.801502 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.89863 (* 0.0909091 = 0.172603 loss) | |
| I0525 04:53:02.801517 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 1.7684 (* 0.0909091 = 0.160764 loss) | |
| I0525 04:53:02.801530 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 1.47611 (* 0.0909091 = 0.134192 loss) | |
| I0525 04:53:02.801543 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.832566 (* 0.0909091 = 0.0756878 loss) | |
| I0525 04:53:02.801558 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.849825 (* 0.0909091 = 0.0772568 loss) | |
| I0525 04:53:02.801571 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.787854 (* 0.0909091 = 0.0716231 loss) | |
| I0525 04:53:02.801585 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.782514 (* 0.0909091 = 0.0711376 loss) | |
| I0525 04:53:02.801599 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.935624 (* 0.0909091 = 0.0850568 loss) | |
| I0525 04:53:02.801614 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.61301 (* 0.0909091 = 0.0557282 loss) | |
| I0525 04:53:02.801627 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 1.00012 (* 0.0909091 = 0.0909201 loss) | |
| I0525 04:53:02.801641 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0091183 (* 0.0909091 = 0.000828937 loss) | |
| I0525 04:53:02.801656 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00363059 (* 0.0909091 = 0.000330054 loss) | |
| I0525 04:53:02.801669 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00321613 (* 0.0909091 = 0.000292375 loss) | |
| I0525 04:53:02.801683 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00123788 (* 0.0909091 = 0.000112534 loss) | |
| I0525 04:53:02.801697 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000677914 (* 0.0909091 = 6.16285e-05 loss) | |
| I0525 04:53:02.801709 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:53:02.801720 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:53:02.801731 5272 solver.cpp:245] Train net output #149: total_confidence = 1.08685e-07 | |
| I0525 04:53:02.801753 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 2.05321e-06 | |
| I0525 04:53:02.801770 5272 sgd_solver.cpp:106] Iteration 21500, lr = 0.001 | |
| I0525 04:57:20.314046 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.9544 > 30) by scale factor 0.938837 | |
| I0525 04:59:23.458575 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 33.0836 > 30) by scale factor 0.906793 | |
| I0525 04:59:27.304214 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.1085 > 30) by scale factor 0.934333 | |
| I0525 04:59:27.710850 5272 solver.cpp:229] Iteration 22000, loss = 9.6773 | |
| I0525 04:59:27.710909 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0677966 | |
| I0525 04:59:27.710928 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0 | |
| I0525 04:59:27.710942 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 04:59:27.710955 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 04:59:27.710968 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0 | |
| I0525 04:59:27.710980 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.125 | |
| I0525 04:59:27.710993 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0 | |
| I0525 04:59:27.711004 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.5 | |
| I0525 04:59:27.711017 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 0.75 | |
| I0525 04:59:27.711030 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 0.875 | |
| I0525 04:59:27.711042 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 0.875 | |
| I0525 04:59:27.711055 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 0.875 | |
| I0525 04:59:27.711067 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 0.875 | |
| I0525 04:59:27.711081 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 0.875 | |
| I0525 04:59:27.711092 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 0.875 | |
| I0525 04:59:27.711104 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 04:59:27.711117 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 04:59:27.711128 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 04:59:27.711140 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 04:59:27.711151 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 04:59:27.711163 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 04:59:27.711175 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 04:59:27.711187 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 04:59:27.711199 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.676136 | |
| I0525 04:59:27.711211 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.20339 | |
| I0525 04:59:27.711227 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.53925 (* 0.3 = 1.06178 loss) | |
| I0525 04:59:27.711241 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 1.31659 (* 0.3 = 0.394978 loss) | |
| I0525 04:59:27.711256 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.38175 (* 0.0272727 = 0.0922296 loss) | |
| I0525 04:59:27.711271 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.57231 (* 0.0272727 = 0.0974266 loss) | |
| I0525 04:59:27.711284 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.68043 (* 0.0272727 = 0.100375 loss) | |
| I0525 04:59:27.711298 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.23006 (* 0.0272727 = 0.0880927 loss) | |
| I0525 04:59:27.711313 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 3.80679 (* 0.0272727 = 0.103821 loss) | |
| I0525 04:59:27.711326 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 3.81571 (* 0.0272727 = 0.104065 loss) | |
| I0525 04:59:27.711340 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 2.41608 (* 0.0272727 = 0.0658932 loss) | |
| I0525 04:59:27.711354 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 1.38785 (* 0.0272727 = 0.0378505 loss) | |
| I0525 04:59:27.711369 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.566975 (* 0.0272727 = 0.0154629 loss) | |
| I0525 04:59:27.711382 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.487841 (* 0.0272727 = 0.0133047 loss) | |
| I0525 04:59:27.711396 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.534627 (* 0.0272727 = 0.0145807 loss) | |
| I0525 04:59:27.711447 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.709563 (* 0.0272727 = 0.0193517 loss) | |
| I0525 04:59:27.711462 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.534905 (* 0.0272727 = 0.0145883 loss) | |
| I0525 04:59:27.711477 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.676085 (* 0.0272727 = 0.0184387 loss) | |
| I0525 04:59:27.711491 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.0711489 (* 0.0272727 = 0.00194042 loss) | |
| I0525 04:59:27.711505 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.035535 (* 0.0272727 = 0.000969137 loss) | |
| I0525 04:59:27.711519 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.01721 (* 0.0272727 = 0.000469364 loss) | |
| I0525 04:59:27.711534 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.0158687 (* 0.0272727 = 0.000432783 loss) | |
| I0525 04:59:27.711547 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.0067653 (* 0.0272727 = 0.000184508 loss) | |
| I0525 04:59:27.711561 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.00304532 (* 0.0272727 = 8.30543e-05 loss) | |
| I0525 04:59:27.711575 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.00573927 (* 0.0272727 = 0.000156525 loss) | |
| I0525 04:59:27.711588 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.00419135 (* 0.0272727 = 0.00011431 loss) | |
| I0525 04:59:27.711601 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0677966 | |
| I0525 04:59:27.711613 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 04:59:27.711626 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 04:59:27.711637 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0.125 | |
| I0525 04:59:27.711649 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0 | |
| I0525 04:59:27.711660 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.125 | |
| I0525 04:59:27.711673 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0 | |
| I0525 04:59:27.711685 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.5 | |
| I0525 04:59:27.711697 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 0.75 | |
| I0525 04:59:27.711709 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 0.875 | |
| I0525 04:59:27.711721 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 0.875 | |
| I0525 04:59:27.711732 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 0.875 | |
| I0525 04:59:27.711745 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 0.875 | |
| I0525 04:59:27.711760 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 0.875 | |
| I0525 04:59:27.711771 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 0.875 | |
| I0525 04:59:27.711783 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 04:59:27.711796 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 04:59:27.711807 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 04:59:27.711818 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 04:59:27.711830 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 04:59:27.711841 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 04:59:27.711853 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 04:59:27.711865 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 04:59:27.711879 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.664773 | |
| I0525 04:59:27.711891 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.152542 | |
| I0525 04:59:27.711906 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.46262 (* 0.3 = 1.03879 loss) | |
| I0525 04:59:27.711920 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 1.35938 (* 0.3 = 0.407815 loss) | |
| I0525 04:59:27.711933 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.13133 (* 0.0272727 = 0.0853998 loss) | |
| I0525 04:59:27.711958 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.49872 (* 0.0272727 = 0.0954195 loss) | |
| I0525 04:59:27.711973 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.31344 (* 0.0272727 = 0.0903664 loss) | |
| I0525 04:59:27.711987 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.63836 (* 0.0272727 = 0.099228 loss) | |
| I0525 04:59:27.712002 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 3.57897 (* 0.0272727 = 0.0976083 loss) | |
| I0525 04:59:27.712014 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 3.81665 (* 0.0272727 = 0.10409 loss) | |
| I0525 04:59:27.712028 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 2.70241 (* 0.0272727 = 0.0737022 loss) | |
| I0525 04:59:27.712041 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 1.241 (* 0.0272727 = 0.0338456 loss) | |
| I0525 04:59:27.712055 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.828625 (* 0.0272727 = 0.0225989 loss) | |
| I0525 04:59:27.712069 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.695958 (* 0.0272727 = 0.0189807 loss) | |
| I0525 04:59:27.712082 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.747721 (* 0.0272727 = 0.0203924 loss) | |
| I0525 04:59:27.712096 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.56435 (* 0.0272727 = 0.0153914 loss) | |
| I0525 04:59:27.712110 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.450707 (* 0.0272727 = 0.012292 loss) | |
| I0525 04:59:27.712124 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.610814 (* 0.0272727 = 0.0166586 loss) | |
| I0525 04:59:27.712138 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.0709232 (* 0.0272727 = 0.00193427 loss) | |
| I0525 04:59:27.712152 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.0515115 (* 0.0272727 = 0.00140486 loss) | |
| I0525 04:59:27.712167 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.0288127 (* 0.0272727 = 0.000785802 loss) | |
| I0525 04:59:27.712180 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.0201693 (* 0.0272727 = 0.000550072 loss) | |
| I0525 04:59:27.712194 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.0165472 (* 0.0272727 = 0.000451287 loss) | |
| I0525 04:59:27.712208 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.00708455 (* 0.0272727 = 0.000193215 loss) | |
| I0525 04:59:27.712221 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.00729991 (* 0.0272727 = 0.000199088 loss) | |
| I0525 04:59:27.712232 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00551168 (* 0.0272727 = 0.000150318 loss) | |
| I0525 04:59:27.712240 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.101695 | |
| I0525 04:59:27.712249 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 04:59:27.712261 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.25 | |
| I0525 04:59:27.712273 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0 | |
| I0525 04:59:27.712285 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0 | |
| I0525 04:59:27.712297 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.125 | |
| I0525 04:59:27.712308 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.25 | |
| I0525 04:59:27.712321 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.5 | |
| I0525 04:59:27.712332 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 0.75 | |
| I0525 04:59:27.712344 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 0.875 | |
| I0525 04:59:27.712357 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 0.875 | |
| I0525 04:59:27.712368 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 0.875 | |
| I0525 04:59:27.712379 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 0.875 | |
| I0525 04:59:27.712391 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 0.875 | |
| I0525 04:59:27.712404 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 0.875 | |
| I0525 04:59:27.712424 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 04:59:27.712437 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 04:59:27.712450 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 04:59:27.712460 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 04:59:27.712472 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 04:59:27.712483 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 04:59:27.712496 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 04:59:27.712507 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 04:59:27.712518 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.676136 | |
| I0525 04:59:27.712530 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.186441 | |
| I0525 04:59:27.712544 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.42441 (* 1 = 3.42441 loss) | |
| I0525 04:59:27.712558 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 1.33442 (* 1 = 1.33442 loss) | |
| I0525 04:59:27.712571 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 2.70014 (* 0.0909091 = 0.245467 loss) | |
| I0525 04:59:27.712585 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.09941 (* 0.0909091 = 0.281764 loss) | |
| I0525 04:59:27.712599 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.70807 (* 0.0909091 = 0.337098 loss) | |
| I0525 04:59:27.712613 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.37623 (* 0.0909091 = 0.30693 loss) | |
| I0525 04:59:27.712627 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 4.11733 (* 0.0909091 = 0.374303 loss) | |
| I0525 04:59:27.712641 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 3.88257 (* 0.0909091 = 0.352961 loss) | |
| I0525 04:59:27.712654 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 2.32389 (* 0.0909091 = 0.211263 loss) | |
| I0525 04:59:27.712668 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 1.15855 (* 0.0909091 = 0.105323 loss) | |
| I0525 04:59:27.712682 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.525891 (* 0.0909091 = 0.0478083 loss) | |
| I0525 04:59:27.712695 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.522159 (* 0.0909091 = 0.047469 loss) | |
| I0525 04:59:27.712709 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.742047 (* 0.0909091 = 0.0674588 loss) | |
| I0525 04:59:27.712723 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.479795 (* 0.0909091 = 0.0436178 loss) | |
| I0525 04:59:27.712736 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.417701 (* 0.0909091 = 0.0379728 loss) | |
| I0525 04:59:27.712750 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.600073 (* 0.0909091 = 0.0545521 loss) | |
| I0525 04:59:27.712764 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.0661752 (* 0.0909091 = 0.00601593 loss) | |
| I0525 04:59:27.712779 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.0216227 (* 0.0909091 = 0.0019657 loss) | |
| I0525 04:59:27.712793 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.0146891 (* 0.0909091 = 0.00133537 loss) | |
| I0525 04:59:27.712810 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.0040037 (* 0.0909091 = 0.000363972 loss) | |
| I0525 04:59:27.712826 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.00456805 (* 0.0909091 = 0.000415277 loss) | |
| I0525 04:59:27.712839 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.00261393 (* 0.0909091 = 0.00023763 loss) | |
| I0525 04:59:27.712853 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.00178649 (* 0.0909091 = 0.000162409 loss) | |
| I0525 04:59:27.712867 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.0012882 (* 0.0909091 = 0.000117109 loss) | |
| I0525 04:59:27.712879 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 04:59:27.712900 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 04:59:27.712913 5272 solver.cpp:245] Train net output #149: total_confidence = 1.89541e-06 | |
| I0525 04:59:27.712927 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 7.59691e-05 | |
| I0525 04:59:27.712942 5272 sgd_solver.cpp:106] Iteration 22000, lr = 0.001 | |
| I0525 05:05:52.771066 5272 solver.cpp:229] Iteration 22500, loss = 9.66745 | |
| I0525 05:05:52.771252 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.025641 | |
| I0525 05:05:52.771275 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 05:05:52.771291 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0 | |
| I0525 05:05:52.771302 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 05:05:52.771316 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 05:05:52.771327 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 05:05:52.771340 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.625 | |
| I0525 05:05:52.771353 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.875 | |
| I0525 05:05:52.771365 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 05:05:52.771378 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 05:05:52.771390 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 05:05:52.771402 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 05:05:52.771415 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 05:05:52.771430 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 05:05:52.771440 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 05:05:52.771452 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 05:05:52.771464 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 05:05:52.771476 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 05:05:52.771488 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 05:05:52.771500 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 05:05:52.771513 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 05:05:52.771524 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 05:05:52.771538 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 05:05:52.771548 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0525 05:05:52.771561 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.128205 | |
| I0525 05:05:52.771579 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 3.39485 (* 0.3 = 1.01846 loss) | |
| I0525 05:05:52.771594 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.858025 (* 0.3 = 0.257408 loss) | |
| I0525 05:05:52.771607 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 3.09767 (* 0.0272727 = 0.0844819 loss) | |
| I0525 05:05:52.771621 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 3.50265 (* 0.0272727 = 0.0955269 loss) | |
| I0525 05:05:52.771636 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 3.30974 (* 0.0272727 = 0.0902655 loss) | |
| I0525 05:05:52.771651 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 3.34645 (* 0.0272727 = 0.0912669 loss) | |
| I0525 05:05:52.771666 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 2.17042 (* 0.0272727 = 0.0591932 loss) | |
| I0525 05:05:52.771679 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.72553 (* 0.0272727 = 0.0470599 loss) | |
| I0525 05:05:52.771693 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 0.743822 (* 0.0272727 = 0.020286 loss) | |
| I0525 05:05:52.771708 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.047318 (* 0.0272727 = 0.00129049 loss) | |
| I0525 05:05:52.771723 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.00618761 (* 0.0272727 = 0.000168753 loss) | |
| I0525 05:05:52.771736 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.00725301 (* 0.0272727 = 0.000197809 loss) | |
| I0525 05:05:52.771751 5272 solver.cpp:245] Train net output #37: loss1/loss11 = 0.00543072 (* 0.0272727 = 0.00014811 loss) | |
| I0525 05:05:52.771765 5272 solver.cpp:245] Train net output #38: loss1/loss12 = 0.00255894 (* 0.0272727 = 6.97892e-05 loss) | |
| I0525 05:05:52.771780 5272 solver.cpp:245] Train net output #39: loss1/loss13 = 0.00233785 (* 0.0272727 = 6.37594e-05 loss) | |
| I0525 05:05:52.771807 5272 solver.cpp:245] Train net output #40: loss1/loss14 = 0.00213642 (* 0.0272727 = 5.82661e-05 loss) | |
| I0525 05:05:52.771822 5272 solver.cpp:245] Train net output #41: loss1/loss15 = 0.00276107 (* 0.0272727 = 7.53018e-05 loss) | |
| I0525 05:05:52.771836 5272 solver.cpp:245] Train net output #42: loss1/loss16 = 0.00191073 (* 0.0272727 = 5.21107e-05 loss) | |
| I0525 05:05:52.771852 5272 solver.cpp:245] Train net output #43: loss1/loss17 = 0.00142779 (* 0.0272727 = 3.89398e-05 loss) | |
| I0525 05:05:52.771865 5272 solver.cpp:245] Train net output #44: loss1/loss18 = 0.00116772 (* 0.0272727 = 3.18468e-05 loss) | |
| I0525 05:05:52.771883 5272 solver.cpp:245] Train net output #45: loss1/loss19 = 0.00112325 (* 0.0272727 = 3.0634e-05 loss) | |
| I0525 05:05:52.771898 5272 solver.cpp:245] Train net output #46: loss1/loss20 = 0.000756945 (* 0.0272727 = 2.0644e-05 loss) | |
| I0525 05:05:52.771911 5272 solver.cpp:245] Train net output #47: loss1/loss21 = 0.000585398 (* 0.0272727 = 1.59654e-05 loss) | |
| I0525 05:05:52.771925 5272 solver.cpp:245] Train net output #48: loss1/loss22 = 0.000817819 (* 0.0272727 = 2.23042e-05 loss) | |
| I0525 05:05:52.771940 5272 solver.cpp:245] Train net output #49: loss2/accuracy = 0.0512821 | |
| I0525 05:05:52.771971 5272 solver.cpp:245] Train net output #50: loss2/accuracy01 = 0.125 | |
| I0525 05:05:52.771984 5272 solver.cpp:245] Train net output #51: loss2/accuracy02 = 0 | |
| I0525 05:05:52.771996 5272 solver.cpp:245] Train net output #52: loss2/accuracy03 = 0 | |
| I0525 05:05:52.772008 5272 solver.cpp:245] Train net output #53: loss2/accuracy04 = 0.25 | |
| I0525 05:05:52.772020 5272 solver.cpp:245] Train net output #54: loss2/accuracy05 = 0.375 | |
| I0525 05:05:52.772033 5272 solver.cpp:245] Train net output #55: loss2/accuracy06 = 0.625 | |
| I0525 05:05:52.772045 5272 solver.cpp:245] Train net output #56: loss2/accuracy07 = 0.875 | |
| I0525 05:05:52.772058 5272 solver.cpp:245] Train net output #57: loss2/accuracy08 = 1 | |
| I0525 05:05:52.772069 5272 solver.cpp:245] Train net output #58: loss2/accuracy09 = 1 | |
| I0525 05:05:52.772081 5272 solver.cpp:245] Train net output #59: loss2/accuracy10 = 1 | |
| I0525 05:05:52.772094 5272 solver.cpp:245] Train net output #60: loss2/accuracy11 = 1 | |
| I0525 05:05:52.772105 5272 solver.cpp:245] Train net output #61: loss2/accuracy12 = 1 | |
| I0525 05:05:52.772116 5272 solver.cpp:245] Train net output #62: loss2/accuracy13 = 1 | |
| I0525 05:05:52.772128 5272 solver.cpp:245] Train net output #63: loss2/accuracy14 = 1 | |
| I0525 05:05:52.772140 5272 solver.cpp:245] Train net output #64: loss2/accuracy15 = 1 | |
| I0525 05:05:52.772152 5272 solver.cpp:245] Train net output #65: loss2/accuracy16 = 1 | |
| I0525 05:05:52.772164 5272 solver.cpp:245] Train net output #66: loss2/accuracy17 = 1 | |
| I0525 05:05:52.772176 5272 solver.cpp:245] Train net output #67: loss2/accuracy18 = 1 | |
| I0525 05:05:52.772187 5272 solver.cpp:245] Train net output #68: loss2/accuracy19 = 1 | |
| I0525 05:05:52.772199 5272 solver.cpp:245] Train net output #69: loss2/accuracy20 = 1 | |
| I0525 05:05:52.772212 5272 solver.cpp:245] Train net output #70: loss2/accuracy21 = 1 | |
| I0525 05:05:52.772223 5272 solver.cpp:245] Train net output #71: loss2/accuracy22 = 1 | |
| I0525 05:05:52.772235 5272 solver.cpp:245] Train net output #72: loss2/accuracy_incl_empty = 0.789773 | |
| I0525 05:05:52.772246 5272 solver.cpp:245] Train net output #73: loss2/accuracy_top3 = 0.153846 | |
| I0525 05:05:52.772266 5272 solver.cpp:245] Train net output #74: loss2/cross_entropy_loss = 3.34064 (* 0.3 = 1.00219 loss) | |
| I0525 05:05:52.772281 5272 solver.cpp:245] Train net output #75: loss2/cross_entropy_loss_incl_empty = 0.795669 (* 0.3 = 0.238701 loss) | |
| I0525 05:05:52.772295 5272 solver.cpp:245] Train net output #76: loss2/loss01 = 3.10648 (* 0.0272727 = 0.0847223 loss) | |
| I0525 05:05:52.772310 5272 solver.cpp:245] Train net output #77: loss2/loss02 = 3.60074 (* 0.0272727 = 0.098202 loss) | |
| I0525 05:05:52.772336 5272 solver.cpp:245] Train net output #78: loss2/loss03 = 3.48501 (* 0.0272727 = 0.0950457 loss) | |
| I0525 05:05:52.772351 5272 solver.cpp:245] Train net output #79: loss2/loss04 = 3.1105 (* 0.0272727 = 0.0848317 loss) | |
| I0525 05:05:52.772364 5272 solver.cpp:245] Train net output #80: loss2/loss05 = 2.10976 (* 0.0272727 = 0.057539 loss) | |
| I0525 05:05:52.772378 5272 solver.cpp:245] Train net output #81: loss2/loss06 = 1.71483 (* 0.0272727 = 0.0467682 loss) | |
| I0525 05:05:52.772392 5272 solver.cpp:245] Train net output #82: loss2/loss07 = 0.717032 (* 0.0272727 = 0.0195554 loss) | |
| I0525 05:05:52.772408 5272 solver.cpp:245] Train net output #83: loss2/loss08 = 0.0208824 (* 0.0272727 = 0.00056952 loss) | |
| I0525 05:05:52.772421 5272 solver.cpp:245] Train net output #84: loss2/loss09 = 0.00340411 (* 0.0272727 = 9.28394e-05 loss) | |
| I0525 05:05:52.772435 5272 solver.cpp:245] Train net output #85: loss2/loss10 = 0.00269401 (* 0.0272727 = 7.3473e-05 loss) | |
| I0525 05:05:52.772450 5272 solver.cpp:245] Train net output #86: loss2/loss11 = 0.00172615 (* 0.0272727 = 4.70767e-05 loss) | |
| I0525 05:05:52.772464 5272 solver.cpp:245] Train net output #87: loss2/loss12 = 0.0015986 (* 0.0272727 = 4.35982e-05 loss) | |
| I0525 05:05:52.772480 5272 solver.cpp:245] Train net output #88: loss2/loss13 = 0.00102282 (* 0.0272727 = 2.7895e-05 loss) | |
| I0525 05:05:52.772493 5272 solver.cpp:245] Train net output #89: loss2/loss14 = 0.00153417 (* 0.0272727 = 4.18409e-05 loss) | |
| I0525 05:05:52.772507 5272 solver.cpp:245] Train net output #90: loss2/loss15 = 0.00158768 (* 0.0272727 = 4.33005e-05 loss) | |
| I0525 05:05:52.772518 5272 solver.cpp:245] Train net output #91: loss2/loss16 = 0.000801379 (* 0.0272727 = 2.18558e-05 loss) | |
| I0525 05:05:52.772528 5272 solver.cpp:245] Train net output #92: loss2/loss17 = 0.00073574 (* 0.0272727 = 2.00656e-05 loss) | |
| I0525 05:05:52.772538 5272 solver.cpp:245] Train net output #93: loss2/loss18 = 0.000514287 (* 0.0272727 = 1.4026e-05 loss) | |
| I0525 05:05:52.772553 5272 solver.cpp:245] Train net output #94: loss2/loss19 = 0.000798053 (* 0.0272727 = 2.17651e-05 loss) | |
| I0525 05:05:52.772567 5272 solver.cpp:245] Train net output #95: loss2/loss20 = 0.000662163 (* 0.0272727 = 1.8059e-05 loss) | |
| I0525 05:05:52.772581 5272 solver.cpp:245] Train net output #96: loss2/loss21 = 0.000623151 (* 0.0272727 = 1.6995e-05 loss) | |
| I0525 05:05:52.772594 5272 solver.cpp:245] Train net output #97: loss2/loss22 = 0.00074224 (* 0.0272727 = 2.02429e-05 loss) | |
| I0525 05:05:52.772608 5272 solver.cpp:245] Train net output #98: loss3/accuracy = 0.0512821 | |
| I0525 05:05:52.772619 5272 solver.cpp:245] Train net output #99: loss3/accuracy01 = 0.125 | |
| I0525 05:05:52.772632 5272 solver.cpp:245] Train net output #100: loss3/accuracy02 = 0.125 | |
| I0525 05:05:52.772644 5272 solver.cpp:245] Train net output #101: loss3/accuracy03 = 0.125 | |
| I0525 05:05:52.772656 5272 solver.cpp:245] Train net output #102: loss3/accuracy04 = 0.25 | |
| I0525 05:05:52.772668 5272 solver.cpp:245] Train net output #103: loss3/accuracy05 = 0.375 | |
| I0525 05:05:52.772680 5272 solver.cpp:245] Train net output #104: loss3/accuracy06 = 0.625 | |
| I0525 05:05:52.772692 5272 solver.cpp:245] Train net output #105: loss3/accuracy07 = 0.875 | |
| I0525 05:05:52.772704 5272 solver.cpp:245] Train net output #106: loss3/accuracy08 = 1 | |
| I0525 05:05:52.772716 5272 solver.cpp:245] Train net output #107: loss3/accuracy09 = 1 | |
| I0525 05:05:52.772728 5272 solver.cpp:245] Train net output #108: loss3/accuracy10 = 1 | |
| I0525 05:05:52.772740 5272 solver.cpp:245] Train net output #109: loss3/accuracy11 = 1 | |
| I0525 05:05:52.772752 5272 solver.cpp:245] Train net output #110: loss3/accuracy12 = 1 | |
| I0525 05:05:52.772763 5272 solver.cpp:245] Train net output #111: loss3/accuracy13 = 1 | |
| I0525 05:05:52.772775 5272 solver.cpp:245] Train net output #112: loss3/accuracy14 = 1 | |
| I0525 05:05:52.772786 5272 solver.cpp:245] Train net output #113: loss3/accuracy15 = 1 | |
| I0525 05:05:52.772799 5272 solver.cpp:245] Train net output #114: loss3/accuracy16 = 1 | |
| I0525 05:05:52.772819 5272 solver.cpp:245] Train net output #115: loss3/accuracy17 = 1 | |
| I0525 05:05:52.772831 5272 solver.cpp:245] Train net output #116: loss3/accuracy18 = 1 | |
| I0525 05:05:52.772843 5272 solver.cpp:245] Train net output #117: loss3/accuracy19 = 1 | |
| I0525 05:05:52.772855 5272 solver.cpp:245] Train net output #118: loss3/accuracy20 = 1 | |
| I0525 05:05:52.772867 5272 solver.cpp:245] Train net output #119: loss3/accuracy21 = 1 | |
| I0525 05:05:52.772879 5272 solver.cpp:245] Train net output #120: loss3/accuracy22 = 1 | |
| I0525 05:05:52.772891 5272 solver.cpp:245] Train net output #121: loss3/accuracy_incl_empty = 0.789773 | |
| I0525 05:05:52.772903 5272 solver.cpp:245] Train net output #122: loss3/accuracy_top3 = 0.282051 | |
| I0525 05:05:52.772917 5272 solver.cpp:245] Train net output #123: loss3/cross_entropy_loss = 3.14293 (* 1 = 3.14293 loss) | |
| I0525 05:05:52.772934 5272 solver.cpp:245] Train net output #124: loss3/cross_entropy_loss_incl_empty = 0.73963 (* 1 = 0.73963 loss) | |
| I0525 05:05:52.772948 5272 solver.cpp:245] Train net output #125: loss3/loss01 = 3.15797 (* 0.0909091 = 0.287088 loss) | |
| I0525 05:05:52.772964 5272 solver.cpp:245] Train net output #126: loss3/loss02 = 3.38184 (* 0.0909091 = 0.30744 loss) | |
| I0525 05:05:52.772977 5272 solver.cpp:245] Train net output #127: loss3/loss03 = 3.0569 (* 0.0909091 = 0.2779 loss) | |
| I0525 05:05:52.772991 5272 solver.cpp:245] Train net output #128: loss3/loss04 = 3.0533 (* 0.0909091 = 0.277573 loss) | |
| I0525 05:05:52.773005 5272 solver.cpp:245] Train net output #129: loss3/loss05 = 2.09699 (* 0.0909091 = 0.190636 loss) | |
| I0525 05:05:52.773020 5272 solver.cpp:245] Train net output #130: loss3/loss06 = 1.79242 (* 0.0909091 = 0.162947 loss) | |
| I0525 05:05:52.773032 5272 solver.cpp:245] Train net output #131: loss3/loss07 = 0.487128 (* 0.0909091 = 0.0442844 loss) | |
| I0525 05:05:52.773047 5272 solver.cpp:245] Train net output #132: loss3/loss08 = 0.0106811 (* 0.0909091 = 0.000971013 loss) | |
| I0525 05:05:52.773061 5272 solver.cpp:245] Train net output #133: loss3/loss09 = 0.00200731 (* 0.0909091 = 0.000182483 loss) | |
| I0525 05:05:52.773075 5272 solver.cpp:245] Train net output #134: loss3/loss10 = 0.00164066 (* 0.0909091 = 0.000149151 loss) | |
| I0525 05:05:52.773089 5272 solver.cpp:245] Train net output #135: loss3/loss11 = 0.00133145 (* 0.0909091 = 0.000121041 loss) | |
| I0525 05:05:52.773103 5272 solver.cpp:245] Train net output #136: loss3/loss12 = 0.0008459 (* 0.0909091 = 7.69e-05 loss) | |
| I0525 05:05:52.773129 5272 solver.cpp:245] Train net output #137: loss3/loss13 = 0.000960478 (* 0.0909091 = 8.73161e-05 loss) | |
| I0525 05:05:52.773147 5272 solver.cpp:245] Train net output #138: loss3/loss14 = 0.000782418 (* 0.0909091 = 7.11289e-05 loss) | |
| I0525 05:05:52.773162 5272 solver.cpp:245] Train net output #139: loss3/loss15 = 0.000765727 (* 0.0909091 = 6.96115e-05 loss) | |
| I0525 05:05:52.773176 5272 solver.cpp:245] Train net output #140: loss3/loss16 = 0.000790156 (* 0.0909091 = 7.18324e-05 loss) | |
| I0525 05:05:52.773190 5272 solver.cpp:245] Train net output #141: loss3/loss17 = 0.000697583 (* 0.0909091 = 6.34167e-05 loss) | |
| I0525 05:05:52.773205 5272 solver.cpp:245] Train net output #142: loss3/loss18 = 0.000769554 (* 0.0909091 = 6.99595e-05 loss) | |
| I0525 05:05:52.773218 5272 solver.cpp:245] Train net output #143: loss3/loss19 = 0.000939737 (* 0.0909091 = 8.54306e-05 loss) | |
| I0525 05:05:52.773232 5272 solver.cpp:245] Train net output #144: loss3/loss20 = 0.000655659 (* 0.0909091 = 5.96054e-05 loss) | |
| I0525 05:05:52.773243 5272 solver.cpp:245] Train net output #145: loss3/loss21 = 0.000603992 (* 0.0909091 = 5.49084e-05 loss) | |
| I0525 05:05:52.773258 5272 solver.cpp:245] Train net output #146: loss3/loss22 = 0.000866675 (* 0.0909091 = 7.87887e-05 loss) | |
| I0525 05:05:52.773272 5272 solver.cpp:245] Train net output #147: total_accuracy = 0 | |
| I0525 05:05:52.773284 5272 solver.cpp:245] Train net output #148: total_accuracy_not_rec = 0 | |
| I0525 05:05:52.773306 5272 solver.cpp:245] Train net output #149: total_confidence = 0.000205044 | |
| I0525 05:05:52.773325 5272 solver.cpp:245] Train net output #150: total_confidence_not_rec = 0.000596169 | |
| I0525 05:05:52.773339 5272 sgd_solver.cpp:106] Iteration 22500, lr = 0.001 | |
| I0525 05:07:33.204154 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 40.8948 > 30) by scale factor 0.73359 | |
| I0525 05:10:22.586864 5272 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 30.8209 > 30) by scale factor 0.973366 | |
| I0525 05:12:17.753682 5272 solver.cpp:229] Iteration 23000, loss = 9.64775 | |
| I0525 05:12:17.753782 5272 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0731707 | |
| I0525 05:12:17.753803 5272 solver.cpp:245] Train net output #1: loss1/accuracy01 = 0.125 | |
| I0525 05:12:17.753815 5272 solver.cpp:245] Train net output #2: loss1/accuracy02 = 0.375 | |
| I0525 05:12:17.753829 5272 solver.cpp:245] Train net output #3: loss1/accuracy03 = 0 | |
| I0525 05:12:17.753841 5272 solver.cpp:245] Train net output #4: loss1/accuracy04 = 0.25 | |
| I0525 05:12:17.753854 5272 solver.cpp:245] Train net output #5: loss1/accuracy05 = 0.375 | |
| I0525 05:12:17.753866 5272 solver.cpp:245] Train net output #6: loss1/accuracy06 = 0.75 | |
| I0525 05:12:17.753878 5272 solver.cpp:245] Train net output #7: loss1/accuracy07 = 0.625 | |
| I0525 05:12:17.753891 5272 solver.cpp:245] Train net output #8: loss1/accuracy08 = 1 | |
| I0525 05:12:17.753904 5272 solver.cpp:245] Train net output #9: loss1/accuracy09 = 1 | |
| I0525 05:12:17.753916 5272 solver.cpp:245] Train net output #10: loss1/accuracy10 = 1 | |
| I0525 05:12:17.753929 5272 solver.cpp:245] Train net output #11: loss1/accuracy11 = 1 | |
| I0525 05:12:17.753942 5272 solver.cpp:245] Train net output #12: loss1/accuracy12 = 1 | |
| I0525 05:12:17.753953 5272 solver.cpp:245] Train net output #13: loss1/accuracy13 = 1 | |
| I0525 05:12:17.753965 5272 solver.cpp:245] Train net output #14: loss1/accuracy14 = 1 | |
| I0525 05:12:17.753978 5272 solver.cpp:245] Train net output #15: loss1/accuracy15 = 1 | |
| I0525 05:12:17.753988 5272 solver.cpp:245] Train net output #16: loss1/accuracy16 = 1 | |
| I0525 05:12:17.754000 5272 solver.cpp:245] Train net output #17: loss1/accuracy17 = 1 | |
| I0525 05:12:17.754012 5272 solver.cpp:245] Train net output #18: loss1/accuracy18 = 1 | |
| I0525 05:12:17.754024 5272 solver.cpp:245] Train net output #19: loss1/accuracy19 = 1 | |
| I0525 05:12:17.754036 5272 solver.cpp:245] Train net output #20: loss1/accuracy20 = 1 | |
| I0525 05:12:17.754047 5272 solver.cpp:245] Train net output #21: loss1/accuracy21 = 1 | |
| I0525 05:12:17.754060 5272 solver.cpp:245] Train net output #22: loss1/accuracy22 = 1 | |
| I0525 05:12:17.754072 5272 solver.cpp:245] Train net output #23: loss1/accuracy_incl_empty = 0.772727 | |
| I0525 05:12:17.754084 5272 solver.cpp:245] Train net output #24: loss1/accuracy_top3 = 0.341463 | |
| I0525 05:12:17.754101 5272 solver.cpp:245] Train net output #25: loss1/cross_entropy_loss = 2.84617 (* 0.3 = 0.85385 loss) | |
| I0525 05:12:17.754115 5272 solver.cpp:245] Train net output #26: loss1/cross_entropy_loss_incl_empty = 0.759457 (* 0.3 = 0.227837 loss) | |
| I0525 05:12:17.754130 5272 solver.cpp:245] Train net output #27: loss1/loss01 = 2.69638 (* 0.0272727 = 0.0735375 loss) | |
| I0525 05:12:17.754144 5272 solver.cpp:245] Train net output #28: loss1/loss02 = 2.48426 (* 0.0272727 = 0.0677525 loss) | |
| I0525 05:12:17.754158 5272 solver.cpp:245] Train net output #29: loss1/loss03 = 2.74282 (* 0.0272727 = 0.0748043 loss) | |
| I0525 05:12:17.754173 5272 solver.cpp:245] Train net output #30: loss1/loss04 = 2.5797 (* 0.0272727 = 0.0703554 loss) | |
| I0525 05:12:17.754187 5272 solver.cpp:245] Train net output #31: loss1/loss05 = 1.91836 (* 0.0272727 = 0.0523188 loss) | |
| I0525 05:12:17.754201 5272 solver.cpp:245] Train net output #32: loss1/loss06 = 1.32071 (* 0.0272727 = 0.0360194 loss) | |
| I0525 05:12:17.754215 5272 solver.cpp:245] Train net output #33: loss1/loss07 = 1.44006 (* 0.0272727 = 0.0392745 loss) | |
| I0525 05:12:17.754230 5272 solver.cpp:245] Train net output #34: loss1/loss08 = 0.0979319 (* 0.0272727 = 0.00267087 loss) | |
| I0525 05:12:17.754245 5272 solver.cpp:245] Train net output #35: loss1/loss09 = 0.0234728 (* 0.0272727 = 0.000640167 loss) | |
| I0525 05:12:17.754258 5272 solver.cpp:245] Train net output #36: loss1/loss10 = 0.0087309 (* 0.0272727 = 0.000238115 loss) | |
| I0525 05:12:17.754273 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment