Last active
April 19, 2016 15:17
-
-
Save stas-sl/d3e20a04fd778617075df4f9d149ee36 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
| I0407 23:54:15.468787 3443 solver.cpp:280] Solving mixed_lstm | |
| I0407 23:54:15.468801 3443 solver.cpp:281] Learning Rate Policy: poly | |
| I0407 23:54:16.245584 3443 solver.cpp:229] Iteration 0, loss = 13.8505 | |
| I0407 23:54:16.245641 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0208333 | |
| I0407 23:54:16.245658 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.00568182 | |
| I0407 23:54:16.245672 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.0416667 | |
| I0407 23:54:16.245688 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 4.34597 (* 0.3 = 1.30379 loss) | |
| I0407 23:54:16.245735 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 4.31917 (* 0.3 = 1.29575 loss) | |
| I0407 23:54:16.245749 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0 | |
| I0407 23:54:16.245761 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0 | |
| I0407 23:54:16.245774 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0 | |
| I0407 23:54:16.245787 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 4.35609 (* 0.3 = 1.30683 loss) | |
| I0407 23:54:16.245801 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 4.45959 (* 0.3 = 1.33788 loss) | |
| I0407 23:54:16.245815 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0 | |
| I0407 23:54:16.245826 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0 | |
| I0407 23:54:16.245837 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.0625 | |
| I0407 23:54:16.245851 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 4.28897 (* 1 = 4.28897 loss) | |
| I0407 23:54:16.245865 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 4.31729 (* 1 = 4.31729 loss) | |
| I0407 23:54:16.245877 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0407 23:54:16.245888 3443 solver.cpp:245] Train net output #16: total_confidence = 4.27234e-37 | |
| I0407 23:54:16.245915 3443 sgd_solver.cpp:106] Iteration 0, lr = 0.01 | |
| I0407 23:54:16.278374 3443 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 32.1432 > 30) by scale factor 0.933324 | |
| I0407 23:59:49.577603 3443 solver.cpp:229] Iteration 500, loss = 8.57124 | |
| I0407 23:59:49.577884 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0227273 | |
| I0407 23:59:49.577906 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0407 23:59:49.577922 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.181818 | |
| I0407 23:59:49.577939 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.70578 (* 0.3 = 1.11174 loss) | |
| I0407 23:59:49.577955 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.30314 (* 0.3 = 0.390943 loss) | |
| I0407 23:59:49.577967 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.113636 | |
| I0407 23:59:49.577980 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0407 23:59:49.577992 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.159091 | |
| I0407 23:59:49.578006 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.73571 (* 0.3 = 1.12071 loss) | |
| I0407 23:59:49.578021 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.22712 (* 0.3 = 0.368136 loss) | |
| I0407 23:59:49.578033 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0681818 | |
| I0407 23:59:49.578045 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.732955 | |
| I0407 23:59:49.578058 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.136364 | |
| I0407 23:59:49.578073 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.47405 (* 1 = 3.47405 loss) | |
| I0407 23:59:49.578088 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.11772 (* 1 = 1.11772 loss) | |
| I0407 23:59:49.578099 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0407 23:59:49.578110 3443 solver.cpp:245] Train net output #16: total_confidence = 1.09867e-06 | |
| I0407 23:59:49.578125 3443 sgd_solver.cpp:106] Iteration 500, lr = 0.00999286 | |
| I0408 00:05:22.958407 3443 solver.cpp:229] Iteration 1000, loss = 7.79891 | |
| I0408 00:05:22.958576 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.037037 | |
| I0408 00:05:22.958597 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.704545 | |
| I0408 00:05:22.958611 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.148148 | |
| I0408 00:05:22.958627 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.98178 (* 0.3 = 1.19453 loss) | |
| I0408 00:05:22.958642 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.30737 (* 0.3 = 0.392212 loss) | |
| I0408 00:05:22.958654 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.037037 | |
| I0408 00:05:22.958667 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.704545 | |
| I0408 00:05:22.958679 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.12963 | |
| I0408 00:05:22.958693 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 4.0223 (* 0.3 = 1.20669 loss) | |
| I0408 00:05:22.958706 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.30932 (* 0.3 = 0.392797 loss) | |
| I0408 00:05:22.958719 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0740741 | |
| I0408 00:05:22.958729 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.710227 | |
| I0408 00:05:22.958741 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.148148 | |
| I0408 00:05:22.958755 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.75606 (* 1 = 3.75606 loss) | |
| I0408 00:05:22.958770 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.22914 (* 1 = 1.22914 loss) | |
| I0408 00:05:22.958781 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:05:22.958792 3443 solver.cpp:245] Train net output #16: total_confidence = 1.49262e-07 | |
| I0408 00:05:22.958806 3443 sgd_solver.cpp:106] Iteration 1000, lr = 0.00998571 | |
| I0408 00:10:56.322994 3443 solver.cpp:229] Iteration 1500, loss = 7.52531 | |
| I0408 00:10:56.323139 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0444444 | |
| I0408 00:10:56.323161 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 00:10:56.323174 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.155556 | |
| I0408 00:10:56.323191 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.74326 (* 0.3 = 1.12298 loss) | |
| I0408 00:10:56.323206 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.13512 (* 0.3 = 0.340536 loss) | |
| I0408 00:10:56.323218 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.111111 | |
| I0408 00:10:56.323231 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.767045 | |
| I0408 00:10:56.323245 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.2 | |
| I0408 00:10:56.323258 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.75473 (* 0.3 = 1.12642 loss) | |
| I0408 00:10:56.323272 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.18927 (* 0.3 = 0.35678 loss) | |
| I0408 00:10:56.323283 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.111111 | |
| I0408 00:10:56.323295 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 00:10:56.323307 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.2 | |
| I0408 00:10:56.323338 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.67668 (* 1 = 3.67668 loss) | |
| I0408 00:10:56.323354 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.973071 (* 1 = 0.973071 loss) | |
| I0408 00:10:56.323366 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:10:56.323379 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000113603 | |
| I0408 00:10:56.323393 3443 sgd_solver.cpp:106] Iteration 1500, lr = 0.00997857 | |
| I0408 00:16:29.714017 3443 solver.cpp:229] Iteration 2000, loss = 7.3746 | |
| I0408 00:16:29.714135 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0408 00:16:29.714154 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 00:16:29.714169 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.148936 | |
| I0408 00:16:29.714184 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.99205 (* 0.3 = 1.19762 loss) | |
| I0408 00:16:29.714200 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.15481 (* 0.3 = 0.346444 loss) | |
| I0408 00:16:29.714213 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0212766 | |
| I0408 00:16:29.714226 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0408 00:16:29.714238 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.106383 | |
| I0408 00:16:29.714252 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 4.06675 (* 0.3 = 1.22002 loss) | |
| I0408 00:16:29.714267 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.22205 (* 0.3 = 0.366616 loss) | |
| I0408 00:16:29.714278 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0638298 | |
| I0408 00:16:29.714290 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.727273 | |
| I0408 00:16:29.714303 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.12766 | |
| I0408 00:16:29.714316 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.54962 (* 1 = 3.54962 loss) | |
| I0408 00:16:29.714329 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.1295 (* 1 = 1.1295 loss) | |
| I0408 00:16:29.714341 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:16:29.714354 3443 solver.cpp:245] Train net output #16: total_confidence = 9.01431e-08 | |
| I0408 00:16:29.714368 3443 sgd_solver.cpp:106] Iteration 2000, lr = 0.00997143 | |
| I0408 00:22:03.105708 3443 solver.cpp:229] Iteration 2500, loss = 7.32302 | |
| I0408 00:22:03.105895 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.111111 | |
| I0408 00:22:03.105933 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 00:22:03.105960 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.244444 | |
| I0408 00:22:03.105993 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.62751 (* 0.3 = 1.08825 loss) | |
| I0408 00:22:03.106025 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.18702 (* 0.3 = 0.356107 loss) | |
| I0408 00:22:03.106051 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.155556 | |
| I0408 00:22:03.106077 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 00:22:03.106103 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.266667 | |
| I0408 00:22:03.106133 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.5063 (* 0.3 = 1.05189 loss) | |
| I0408 00:22:03.106163 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.15554 (* 0.3 = 0.346663 loss) | |
| I0408 00:22:03.106189 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.133333 | |
| I0408 00:22:03.106215 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 00:22:03.106238 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.222222 | |
| I0408 00:22:03.106267 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.38337 (* 1 = 3.38337 loss) | |
| I0408 00:22:03.106294 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.98052 (* 1 = 0.98052 loss) | |
| I0408 00:22:03.106318 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:22:03.106341 3443 solver.cpp:245] Train net output #16: total_confidence = 1.35487e-06 | |
| I0408 00:22:03.106366 3443 sgd_solver.cpp:106] Iteration 2500, lr = 0.00996429 | |
| I0408 00:27:36.487051 3443 solver.cpp:229] Iteration 3000, loss = 7.2269 | |
| I0408 00:27:36.487251 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0576923 | |
| I0408 00:27:36.487272 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.721591 | |
| I0408 00:27:36.487285 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.153846 | |
| I0408 00:27:36.487301 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.4688 (* 0.3 = 1.04064 loss) | |
| I0408 00:27:36.487316 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.1772 (* 0.3 = 0.35316 loss) | |
| I0408 00:27:36.487329 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0961538 | |
| I0408 00:27:36.487341 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.727273 | |
| I0408 00:27:36.487354 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.134615 | |
| I0408 00:27:36.487380 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.67985 (* 0.3 = 1.10396 loss) | |
| I0408 00:27:36.487397 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.2238 (* 0.3 = 0.36714 loss) | |
| I0408 00:27:36.487411 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0576923 | |
| I0408 00:27:36.487423 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.721591 | |
| I0408 00:27:36.487437 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.115385 | |
| I0408 00:27:36.487450 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.40934 (* 1 = 3.40934 loss) | |
| I0408 00:27:36.487464 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.05973 (* 1 = 1.05973 loss) | |
| I0408 00:27:36.487476 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:27:36.487488 3443 solver.cpp:245] Train net output #16: total_confidence = 1.15075e-05 | |
| I0408 00:27:36.487504 3443 sgd_solver.cpp:106] Iteration 3000, lr = 0.00995714 | |
| I0408 00:33:09.884913 3443 solver.cpp:229] Iteration 3500, loss = 7.13741 | |
| I0408 00:33:09.885085 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.06 | |
| I0408 00:33:09.885107 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 00:33:09.885120 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.1 | |
| I0408 00:33:09.885138 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.64733 (* 0.3 = 1.0942 loss) | |
| I0408 00:33:09.885152 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.15128 (* 0.3 = 0.345384 loss) | |
| I0408 00:33:09.885164 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.02 | |
| I0408 00:33:09.885177 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.710227 | |
| I0408 00:33:09.885190 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.1 | |
| I0408 00:33:09.885203 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.79749 (* 0.3 = 1.13925 loss) | |
| I0408 00:33:09.885217 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.2949 (* 0.3 = 0.38847 loss) | |
| I0408 00:33:09.885229 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.06 | |
| I0408 00:33:09.885241 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.721591 | |
| I0408 00:33:09.885253 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.2 | |
| I0408 00:33:09.885268 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.37589 (* 1 = 3.37589 loss) | |
| I0408 00:33:09.885282 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.06474 (* 1 = 1.06474 loss) | |
| I0408 00:33:09.885294 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:33:09.885306 3443 solver.cpp:245] Train net output #16: total_confidence = 3.58602e-07 | |
| I0408 00:33:09.885321 3443 sgd_solver.cpp:106] Iteration 3500, lr = 0.00995 | |
| I0408 00:38:43.268569 3443 solver.cpp:229] Iteration 4000, loss = 7.03864 | |
| I0408 00:38:43.268792 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0612245 | |
| I0408 00:38:43.268813 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 00:38:43.268827 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.183673 | |
| I0408 00:38:43.268844 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.29901 (* 0.3 = 0.989703 loss) | |
| I0408 00:38:43.268858 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.05602 (* 0.3 = 0.316805 loss) | |
| I0408 00:38:43.268872 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0408163 | |
| I0408 00:38:43.268883 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.727273 | |
| I0408 00:38:43.268895 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.22449 | |
| I0408 00:38:43.268909 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.43611 (* 0.3 = 1.03083 loss) | |
| I0408 00:38:43.268926 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.05983 (* 0.3 = 0.31795 loss) | |
| I0408 00:38:43.268939 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0612245 | |
| I0408 00:38:43.268957 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.738636 | |
| I0408 00:38:43.268970 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.204082 | |
| I0408 00:38:43.268985 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.29647 (* 1 = 3.29647 loss) | |
| I0408 00:38:43.268998 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.977883 (* 1 = 0.977883 loss) | |
| I0408 00:38:43.269011 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:38:43.269022 3443 solver.cpp:245] Train net output #16: total_confidence = 6.47519e-06 | |
| I0408 00:38:43.269037 3443 sgd_solver.cpp:106] Iteration 4000, lr = 0.00994286 | |
| I0408 00:44:16.674360 3443 solver.cpp:229] Iteration 4500, loss = 7.02923 | |
| I0408 00:44:16.674512 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.075 | |
| I0408 00:44:16.674533 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 00:44:16.674546 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.15 | |
| I0408 00:44:16.674563 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.36885 (* 0.3 = 1.01065 loss) | |
| I0408 00:44:16.674577 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.12503 (* 0.3 = 0.33751 loss) | |
| I0408 00:44:16.674589 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.125 | |
| I0408 00:44:16.674602 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 00:44:16.674614 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.3 | |
| I0408 00:44:16.674628 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.1542 (* 0.3 = 0.94626 loss) | |
| I0408 00:44:16.674641 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.892524 (* 0.3 = 0.267757 loss) | |
| I0408 00:44:16.674654 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.125 | |
| I0408 00:44:16.674664 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.767045 | |
| I0408 00:44:16.674676 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.325 | |
| I0408 00:44:16.674690 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.05293 (* 1 = 3.05293 loss) | |
| I0408 00:44:16.674705 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.859965 (* 1 = 0.859965 loss) | |
| I0408 00:44:16.674716 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:44:16.674727 3443 solver.cpp:245] Train net output #16: total_confidence = 9.64655e-05 | |
| I0408 00:44:16.674742 3443 sgd_solver.cpp:106] Iteration 4500, lr = 0.00993571 | |
| I0408 00:49:50.311938 3443 solver.cpp:338] Iteration 5000, Testing net (#0) | |
| I0408 00:50:31.495959 3443 solver.cpp:393] Test loss: 8.68262 | |
| I0408 00:50:31.496080 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.0452167 | |
| I0408 00:50:31.496100 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.76641 | |
| I0408 00:50:31.496114 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.119341 | |
| I0408 00:50:31.496130 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 4.58886 (* 0.3 = 1.37666 loss) | |
| I0408 00:50:31.496145 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 1.14863 (* 0.3 = 0.344589 loss) | |
| I0408 00:50:31.496156 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.0387071 | |
| I0408 00:50:31.496170 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.764682 | |
| I0408 00:50:31.496181 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.112384 | |
| I0408 00:50:31.496194 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 4.45574 (* 0.3 = 1.33672 loss) | |
| I0408 00:50:31.496207 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 1.11161 (* 0.3 = 0.333483 loss) | |
| I0408 00:50:31.496219 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.0388385 | |
| I0408 00:50:31.496232 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.761181 | |
| I0408 00:50:31.496243 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.103648 | |
| I0408 00:50:31.496256 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 4.18329 (* 1 = 4.18329 loss) | |
| I0408 00:50:31.496269 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 1.10788 (* 1 = 1.10788 loss) | |
| I0408 00:50:31.496280 3443 solver.cpp:406] Test net output #15: total_accuracy = 0 | |
| I0408 00:50:31.496292 3443 solver.cpp:406] Test net output #16: total_confidence = 2.95794e-05 | |
| I0408 00:50:31.868674 3443 solver.cpp:229] Iteration 5000, loss = 6.96566 | |
| I0408 00:50:31.868736 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0465116 | |
| I0408 00:50:31.868755 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 00:50:31.868768 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.162791 | |
| I0408 00:50:31.868784 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.50065 (* 0.3 = 1.0502 loss) | |
| I0408 00:50:31.868799 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.06964 (* 0.3 = 0.320891 loss) | |
| I0408 00:50:31.868811 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0697674 | |
| I0408 00:50:31.868824 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0408 00:50:31.868836 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.186047 | |
| I0408 00:50:31.868849 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.57923 (* 0.3 = 1.07377 loss) | |
| I0408 00:50:31.868863 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.09333 (* 0.3 = 0.327999 loss) | |
| I0408 00:50:31.868875 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0232558 | |
| I0408 00:50:31.868888 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.744318 | |
| I0408 00:50:31.868901 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.116279 | |
| I0408 00:50:31.868914 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.29118 (* 1 = 3.29118 loss) | |
| I0408 00:50:31.868928 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.934273 (* 1 = 0.934273 loss) | |
| I0408 00:50:31.868940 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:50:31.868952 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00020139 | |
| I0408 00:50:31.868968 3443 sgd_solver.cpp:106] Iteration 5000, lr = 0.00992857 | |
| I0408 00:56:05.103965 3443 solver.cpp:229] Iteration 5500, loss = 6.96715 | |
| I0408 00:56:05.104156 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0625 | |
| I0408 00:56:05.104176 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 00:56:05.104190 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.145833 | |
| I0408 00:56:05.104207 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.38839 (* 0.3 = 1.01652 loss) | |
| I0408 00:56:05.104221 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.97679 (* 0.3 = 0.293037 loss) | |
| I0408 00:56:05.104234 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0208333 | |
| I0408 00:56:05.104246 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.732955 | |
| I0408 00:56:05.104259 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.208333 | |
| I0408 00:56:05.104274 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.48493 (* 0.3 = 1.04548 loss) | |
| I0408 00:56:05.104287 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.02138 (* 0.3 = 0.306415 loss) | |
| I0408 00:56:05.104300 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0416667 | |
| I0408 00:56:05.104312 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.732955 | |
| I0408 00:56:05.104324 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.1875 | |
| I0408 00:56:05.104338 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.38987 (* 1 = 3.38987 loss) | |
| I0408 00:56:05.104352 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.988944 (* 1 = 0.988944 loss) | |
| I0408 00:56:05.104364 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 00:56:05.104377 3443 solver.cpp:245] Train net output #16: total_confidence = 9.83636e-07 | |
| I0408 00:56:05.104392 3443 sgd_solver.cpp:106] Iteration 5500, lr = 0.00992143 | |
| I0408 01:01:38.518291 3443 solver.cpp:229] Iteration 6000, loss = 6.88576 | |
| I0408 01:01:38.518440 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0869565 | |
| I0408 01:01:38.518461 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 01:01:38.518476 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.217391 | |
| I0408 01:01:38.518492 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.35048 (* 0.3 = 1.00514 loss) | |
| I0408 01:01:38.518507 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.01227 (* 0.3 = 0.303681 loss) | |
| I0408 01:01:38.518519 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0434783 | |
| I0408 01:01:38.518532 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.744318 | |
| I0408 01:01:38.518544 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.173913 | |
| I0408 01:01:38.518558 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.42432 (* 0.3 = 1.0273 loss) | |
| I0408 01:01:38.518571 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.04517 (* 0.3 = 0.313551 loss) | |
| I0408 01:01:38.518584 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0869565 | |
| I0408 01:01:38.518595 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.755682 | |
| I0408 01:01:38.518607 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.23913 | |
| I0408 01:01:38.518620 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.23763 (* 1 = 3.23763 loss) | |
| I0408 01:01:38.518635 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.946734 (* 1 = 0.946734 loss) | |
| I0408 01:01:38.518646 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:01:38.518658 3443 solver.cpp:245] Train net output #16: total_confidence = 1.10272e-06 | |
| I0408 01:01:38.518673 3443 sgd_solver.cpp:106] Iteration 6000, lr = 0.00991429 | |
| I0408 01:07:11.878407 3443 solver.cpp:229] Iteration 6500, loss = 6.83744 | |
| I0408 01:07:11.878624 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1 | |
| I0408 01:07:11.878646 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 01:07:11.878659 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.24 | |
| I0408 01:07:11.878676 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.52527 (* 0.3 = 1.05758 loss) | |
| I0408 01:07:11.878691 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.07446 (* 0.3 = 0.322339 loss) | |
| I0408 01:07:11.878705 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.1 | |
| I0408 01:07:11.878716 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.732955 | |
| I0408 01:07:11.878728 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.24 | |
| I0408 01:07:11.878743 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.24729 (* 0.3 = 0.974188 loss) | |
| I0408 01:07:11.878757 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.09492 (* 0.3 = 0.328477 loss) | |
| I0408 01:07:11.878769 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.08 | |
| I0408 01:07:11.878782 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.738636 | |
| I0408 01:07:11.878794 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.3 | |
| I0408 01:07:11.878808 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.24869 (* 1 = 3.24869 loss) | |
| I0408 01:07:11.878823 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.963159 (* 1 = 0.963159 loss) | |
| I0408 01:07:11.878834 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:07:11.878845 3443 solver.cpp:245] Train net output #16: total_confidence = 6.6002e-06 | |
| I0408 01:07:11.878861 3443 sgd_solver.cpp:106] Iteration 6500, lr = 0.00990714 | |
| I0408 01:12:45.268910 3443 solver.cpp:229] Iteration 7000, loss = 6.78223 | |
| I0408 01:12:45.269050 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0465116 | |
| I0408 01:12:45.269070 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 01:12:45.269083 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.255814 | |
| I0408 01:12:45.269109 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.36986 (* 0.3 = 1.01096 loss) | |
| I0408 01:12:45.269124 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.923711 (* 0.3 = 0.277113 loss) | |
| I0408 01:12:45.269136 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0930233 | |
| I0408 01:12:45.269150 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 01:12:45.269161 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.255814 | |
| I0408 01:12:45.269176 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.21002 (* 0.3 = 0.963007 loss) | |
| I0408 01:12:45.269191 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.971047 (* 0.3 = 0.291314 loss) | |
| I0408 01:12:45.269202 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0465116 | |
| I0408 01:12:45.269214 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.761364 | |
| I0408 01:12:45.269227 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.255814 | |
| I0408 01:12:45.269240 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.0852 (* 1 = 3.0852 loss) | |
| I0408 01:12:45.269253 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.821484 (* 1 = 0.821484 loss) | |
| I0408 01:12:45.269265 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:12:45.269278 3443 solver.cpp:245] Train net output #16: total_confidence = 8.29975e-05 | |
| I0408 01:12:45.269291 3443 sgd_solver.cpp:106] Iteration 7000, lr = 0.0099 | |
| I0408 01:18:18.755538 3443 solver.cpp:229] Iteration 7500, loss = 6.75431 | |
| I0408 01:18:18.755767 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0869565 | |
| I0408 01:18:18.755795 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 01:18:18.755808 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.26087 | |
| I0408 01:18:18.755826 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.16702 (* 0.3 = 0.950106 loss) | |
| I0408 01:18:18.755841 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.947802 (* 0.3 = 0.284341 loss) | |
| I0408 01:18:18.755853 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0869565 | |
| I0408 01:18:18.755867 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.75 | |
| I0408 01:18:18.755879 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.152174 | |
| I0408 01:18:18.755903 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.17755 (* 0.3 = 0.953266 loss) | |
| I0408 01:18:18.755916 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.908958 (* 0.3 = 0.272687 loss) | |
| I0408 01:18:18.755933 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.130435 | |
| I0408 01:18:18.755945 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.767045 | |
| I0408 01:18:18.755957 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.26087 | |
| I0408 01:18:18.755971 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.00633 (* 1 = 3.00633 loss) | |
| I0408 01:18:18.755985 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.897838 (* 1 = 0.897838 loss) | |
| I0408 01:18:18.755998 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:18:18.756011 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000113947 | |
| I0408 01:18:18.756026 3443 sgd_solver.cpp:106] Iteration 7500, lr = 0.00989286 | |
| I0408 01:23:52.054919 3443 solver.cpp:229] Iteration 8000, loss = 6.70293 | |
| I0408 01:23:52.055099 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.08 | |
| I0408 01:23:52.055131 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.721591 | |
| I0408 01:23:52.055148 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.12 | |
| I0408 01:23:52.055166 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.30661 (* 0.3 = 0.991982 loss) | |
| I0408 01:23:52.055181 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.0523 (* 0.3 = 0.31569 loss) | |
| I0408 01:23:52.055193 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.04 | |
| I0408 01:23:52.055207 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.727273 | |
| I0408 01:23:52.055218 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.1 | |
| I0408 01:23:52.055233 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.37371 (* 0.3 = 1.01211 loss) | |
| I0408 01:23:52.055246 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.04192 (* 0.3 = 0.312577 loss) | |
| I0408 01:23:52.055258 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.06 | |
| I0408 01:23:52.055270 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.727273 | |
| I0408 01:23:52.055282 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.22 | |
| I0408 01:23:52.055296 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.34 (* 1 = 3.34 loss) | |
| I0408 01:23:52.055310 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.0034 (* 1 = 1.0034 loss) | |
| I0408 01:23:52.055341 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:23:52.055356 3443 solver.cpp:245] Train net output #16: total_confidence = 7.26029e-05 | |
| I0408 01:23:52.055371 3443 sgd_solver.cpp:106] Iteration 8000, lr = 0.00988571 | |
| I0408 01:29:25.509419 3443 solver.cpp:229] Iteration 8500, loss = 6.7423 | |
| I0408 01:29:25.509620 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0638298 | |
| I0408 01:29:25.509642 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 01:29:25.509655 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.191489 | |
| I0408 01:29:25.509672 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.15295 (* 0.3 = 0.945884 loss) | |
| I0408 01:29:25.509687 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.950862 (* 0.3 = 0.285259 loss) | |
| I0408 01:29:25.509701 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.170213 | |
| I0408 01:29:25.509712 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 01:29:25.509724 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.361702 | |
| I0408 01:29:25.509738 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.03076 (* 0.3 = 0.909227 loss) | |
| I0408 01:29:25.509752 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.888856 (* 0.3 = 0.266657 loss) | |
| I0408 01:29:25.509764 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.148936 | |
| I0408 01:29:25.509776 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.761364 | |
| I0408 01:29:25.509788 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.319149 | |
| I0408 01:29:25.509802 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.88177 (* 1 = 2.88177 loss) | |
| I0408 01:29:25.509816 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.875165 (* 1 = 0.875165 loss) | |
| I0408 01:29:25.509829 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:29:25.509840 3443 solver.cpp:245] Train net output #16: total_confidence = 8.48266e-05 | |
| I0408 01:29:25.509855 3443 sgd_solver.cpp:106] Iteration 8500, lr = 0.00987857 | |
| I0408 01:34:59.203960 3443 solver.cpp:229] Iteration 9000, loss = 6.70921 | |
| I0408 01:34:59.204139 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0833333 | |
| I0408 01:34:59.204160 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 01:34:59.204175 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.229167 | |
| I0408 01:34:59.204192 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.23136 (* 0.3 = 0.969409 loss) | |
| I0408 01:34:59.204206 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.981896 (* 0.3 = 0.294569 loss) | |
| I0408 01:34:59.204218 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0833333 | |
| I0408 01:34:59.204231 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.744318 | |
| I0408 01:34:59.204243 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.25 | |
| I0408 01:34:59.204257 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.15698 (* 0.3 = 0.947095 loss) | |
| I0408 01:34:59.204272 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.949926 (* 0.3 = 0.284978 loss) | |
| I0408 01:34:59.204283 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0625 | |
| I0408 01:34:59.204295 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.744318 | |
| I0408 01:34:59.204308 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.208333 | |
| I0408 01:34:59.204321 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.06259 (* 1 = 3.06259 loss) | |
| I0408 01:34:59.204335 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.882237 (* 1 = 0.882237 loss) | |
| I0408 01:34:59.204347 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:34:59.204360 3443 solver.cpp:245] Train net output #16: total_confidence = 2.68575e-05 | |
| I0408 01:34:59.204375 3443 sgd_solver.cpp:106] Iteration 9000, lr = 0.00987143 | |
| I0408 01:40:32.952997 3443 solver.cpp:229] Iteration 9500, loss = 6.63286 | |
| I0408 01:40:32.953204 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0612245 | |
| I0408 01:40:32.953227 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 01:40:32.953240 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.306122 | |
| I0408 01:40:32.953258 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.16729 (* 0.3 = 0.950187 loss) | |
| I0408 01:40:32.953272 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.955458 (* 0.3 = 0.286638 loss) | |
| I0408 01:40:32.953285 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0816327 | |
| I0408 01:40:32.953299 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.727273 | |
| I0408 01:40:32.953310 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.265306 | |
| I0408 01:40:32.953325 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.23331 (* 0.3 = 0.969995 loss) | |
| I0408 01:40:32.953338 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.02186 (* 0.3 = 0.306558 loss) | |
| I0408 01:40:32.953351 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.122449 | |
| I0408 01:40:32.953363 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.75 | |
| I0408 01:40:32.953375 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.244898 | |
| I0408 01:40:32.953390 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.10136 (* 1 = 3.10136 loss) | |
| I0408 01:40:32.953404 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.93279 (* 1 = 0.93279 loss) | |
| I0408 01:40:32.953416 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:40:32.953428 3443 solver.cpp:245] Train net output #16: total_confidence = 6.09814e-07 | |
| I0408 01:40:32.953444 3443 sgd_solver.cpp:106] Iteration 9500, lr = 0.00986429 | |
| I0408 01:46:05.866292 3443 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_10000.caffemodel | |
| I0408 01:46:06.445940 3443 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_10000.solverstate | |
| I0408 01:46:06.698824 3443 solver.cpp:338] Iteration 10000, Testing net (#0) | |
| I0408 01:46:47.708153 3443 solver.cpp:393] Test loss: 6.41976 | |
| I0408 01:46:47.708286 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.115397 | |
| I0408 01:46:47.708305 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.782273 | |
| I0408 01:46:47.708319 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.316424 | |
| I0408 01:46:47.708336 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 3.17685 (* 0.3 = 0.953054 loss) | |
| I0408 01:46:47.708353 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.800331 (* 0.3 = 0.240099 loss) | |
| I0408 01:46:47.708365 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.0954313 | |
| I0408 01:46:47.708379 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.778727 | |
| I0408 01:46:47.708391 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.290776 | |
| I0408 01:46:47.708406 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 3.20138 (* 0.3 = 0.960414 loss) | |
| I0408 01:46:47.708420 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.798825 (* 0.3 = 0.239647 loss) | |
| I0408 01:46:47.708433 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.114487 | |
| I0408 01:46:47.708446 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.779909 | |
| I0408 01:46:47.708457 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.291609 | |
| I0408 01:46:47.708472 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 3.20638 (* 1 = 3.20638 loss) | |
| I0408 01:46:47.708485 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.82017 (* 1 = 0.82017 loss) | |
| I0408 01:46:47.708498 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.002 | |
| I0408 01:46:47.708510 3443 solver.cpp:406] Test net output #16: total_confidence = 0.000754739 | |
| I0408 01:46:48.081652 3443 solver.cpp:229] Iteration 10000, loss = 6.63282 | |
| I0408 01:46:48.081728 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0714286 | |
| I0408 01:46:48.081748 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 01:46:48.081763 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.261905 | |
| I0408 01:46:48.081779 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.15879 (* 0.3 = 0.947638 loss) | |
| I0408 01:46:48.081795 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.944388 (* 0.3 = 0.283316 loss) | |
| I0408 01:46:48.081809 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0714286 | |
| I0408 01:46:48.081821 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.755682 | |
| I0408 01:46:48.081835 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.261905 | |
| I0408 01:46:48.081850 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.99565 (* 0.3 = 0.898696 loss) | |
| I0408 01:46:48.081864 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.888492 (* 0.3 = 0.266548 loss) | |
| I0408 01:46:48.081877 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.119048 | |
| I0408 01:46:48.081890 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.761364 | |
| I0408 01:46:48.081903 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.404762 | |
| I0408 01:46:48.081918 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.99631 (* 1 = 2.99631 loss) | |
| I0408 01:46:48.081933 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.892211 (* 1 = 0.892211 loss) | |
| I0408 01:46:48.081945 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:46:48.081959 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000626543 | |
| I0408 01:46:48.081974 3443 sgd_solver.cpp:106] Iteration 10000, lr = 0.00985714 | |
| I0408 01:52:21.322742 3443 solver.cpp:229] Iteration 10500, loss = 6.60527 | |
| I0408 01:52:21.322854 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0952381 | |
| I0408 01:52:21.322876 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 01:52:21.322890 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.214286 | |
| I0408 01:52:21.322906 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.61797 (* 0.3 = 1.08539 loss) | |
| I0408 01:52:21.322924 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.998391 (* 0.3 = 0.299517 loss) | |
| I0408 01:52:21.322937 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0714286 | |
| I0408 01:52:21.322950 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 01:52:21.322963 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.166667 | |
| I0408 01:52:21.322978 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.65355 (* 0.3 = 1.09606 loss) | |
| I0408 01:52:21.322993 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.959195 (* 0.3 = 0.287759 loss) | |
| I0408 01:52:21.323006 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.119048 | |
| I0408 01:52:21.323019 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.778409 | |
| I0408 01:52:21.323031 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.261905 | |
| I0408 01:52:21.323045 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.53152 (* 1 = 3.53152 loss) | |
| I0408 01:52:21.323060 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.94171 (* 1 = 0.94171 loss) | |
| I0408 01:52:21.323072 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:52:21.323084 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000155565 | |
| I0408 01:52:21.323099 3443 sgd_solver.cpp:106] Iteration 10500, lr = 0.00985 | |
| I0408 01:57:54.694224 3443 solver.cpp:229] Iteration 11000, loss = 6.58077 | |
| I0408 01:57:54.694381 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0416667 | |
| I0408 01:57:54.694402 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.727273 | |
| I0408 01:57:54.694416 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.270833 | |
| I0408 01:57:54.694433 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.18979 (* 0.3 = 0.956937 loss) | |
| I0408 01:57:54.694448 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.990983 (* 0.3 = 0.297295 loss) | |
| I0408 01:57:54.694461 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0833333 | |
| I0408 01:57:54.694474 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0408 01:57:54.694486 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.270833 | |
| I0408 01:57:54.694500 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.17816 (* 0.3 = 0.953449 loss) | |
| I0408 01:57:54.694515 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.959466 (* 0.3 = 0.28784 loss) | |
| I0408 01:57:54.694535 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.104167 | |
| I0408 01:57:54.694547 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.744318 | |
| I0408 01:57:54.694560 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.25 | |
| I0408 01:57:54.694574 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.09052 (* 1 = 3.09052 loss) | |
| I0408 01:57:54.694597 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.944494 (* 1 = 0.944494 loss) | |
| I0408 01:57:54.694609 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 01:57:54.694622 3443 solver.cpp:245] Train net output #16: total_confidence = 2.03084e-05 | |
| I0408 01:57:54.694636 3443 sgd_solver.cpp:106] Iteration 11000, lr = 0.00984286 | |
| I0408 02:03:28.141132 3443 solver.cpp:229] Iteration 11500, loss = 6.53889 | |
| I0408 02:03:28.141249 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0566038 | |
| I0408 02:03:28.141268 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.715909 | |
| I0408 02:03:28.141283 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.188679 | |
| I0408 02:03:28.141299 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.27794 (* 0.3 = 0.983382 loss) | |
| I0408 02:03:28.141315 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.04266 (* 0.3 = 0.312799 loss) | |
| I0408 02:03:28.141329 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0566038 | |
| I0408 02:03:28.141340 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.715909 | |
| I0408 02:03:28.141352 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.245283 | |
| I0408 02:03:28.141367 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.26085 (* 0.3 = 0.978256 loss) | |
| I0408 02:03:28.141381 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.03133 (* 0.3 = 0.309398 loss) | |
| I0408 02:03:28.141394 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0566038 | |
| I0408 02:03:28.141407 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.715909 | |
| I0408 02:03:28.141419 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.245283 | |
| I0408 02:03:28.141433 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.18574 (* 1 = 3.18574 loss) | |
| I0408 02:03:28.141448 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.996939 (* 1 = 0.996939 loss) | |
| I0408 02:03:28.141461 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:03:28.141474 3443 solver.cpp:245] Train net output #16: total_confidence = 6.79727e-06 | |
| I0408 02:03:28.141489 3443 sgd_solver.cpp:106] Iteration 11500, lr = 0.00983571 | |
| I0408 02:09:01.466855 3443 solver.cpp:229] Iteration 12000, loss = 6.58063 | |
| I0408 02:09:01.466995 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.107143 | |
| I0408 02:09:01.467015 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.698864 | |
| I0408 02:09:01.467030 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.214286 | |
| I0408 02:09:01.467046 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.26884 (* 0.3 = 0.980652 loss) | |
| I0408 02:09:01.467062 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.17759 (* 0.3 = 0.353277 loss) | |
| I0408 02:09:01.467075 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0892857 | |
| I0408 02:09:01.467087 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.704545 | |
| I0408 02:09:01.467100 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.285714 | |
| I0408 02:09:01.467114 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.20895 (* 0.3 = 0.962686 loss) | |
| I0408 02:09:01.467129 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.09683 (* 0.3 = 0.329048 loss) | |
| I0408 02:09:01.467142 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.107143 | |
| I0408 02:09:01.467155 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.715909 | |
| I0408 02:09:01.467167 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.267857 | |
| I0408 02:09:01.467182 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.10934 (* 1 = 3.10934 loss) | |
| I0408 02:09:01.467196 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.0198 (* 1 = 1.0198 loss) | |
| I0408 02:09:01.467209 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:09:01.467221 3443 solver.cpp:245] Train net output #16: total_confidence = 1.90635e-05 | |
| I0408 02:09:01.467236 3443 sgd_solver.cpp:106] Iteration 12000, lr = 0.00982857 | |
| I0408 02:14:34.832674 3443 solver.cpp:229] Iteration 12500, loss = 6.47738 | |
| I0408 02:14:34.832792 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.102041 | |
| I0408 02:14:34.832811 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 02:14:34.832825 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.183673 | |
| I0408 02:14:34.832842 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.32352 (* 0.3 = 0.997055 loss) | |
| I0408 02:14:34.832859 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.981483 (* 0.3 = 0.294445 loss) | |
| I0408 02:14:34.832870 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0816327 | |
| I0408 02:14:34.832883 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.744318 | |
| I0408 02:14:34.832896 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.204082 | |
| I0408 02:14:34.832911 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.3199 (* 0.3 = 0.99597 loss) | |
| I0408 02:14:34.832926 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.971394 (* 0.3 = 0.291418 loss) | |
| I0408 02:14:34.832937 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.102041 | |
| I0408 02:14:34.832954 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.75 | |
| I0408 02:14:34.832968 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.204082 | |
| I0408 02:14:34.832983 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.31281 (* 1 = 3.31281 loss) | |
| I0408 02:14:34.832998 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.992899 (* 1 = 0.992899 loss) | |
| I0408 02:14:34.833010 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:14:34.833024 3443 solver.cpp:245] Train net output #16: total_confidence = 6.83323e-05 | |
| I0408 02:14:34.833039 3443 sgd_solver.cpp:106] Iteration 12500, lr = 0.00982143 | |
| I0408 02:20:08.237329 3443 solver.cpp:229] Iteration 13000, loss = 6.4472 | |
| I0408 02:20:08.237529 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0 | |
| I0408 02:20:08.237551 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.715909 | |
| I0408 02:20:08.237566 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.12766 | |
| I0408 02:20:08.237584 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 4.31029 (* 0.3 = 1.29309 loss) | |
| I0408 02:20:08.237599 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.25873 (* 0.3 = 0.377619 loss) | |
| I0408 02:20:08.237612 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.106383 | |
| I0408 02:20:08.237627 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.75 | |
| I0408 02:20:08.237639 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.191489 | |
| I0408 02:20:08.237654 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 4.31412 (* 0.3 = 1.29423 loss) | |
| I0408 02:20:08.237669 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.23987 (* 0.3 = 0.37196 loss) | |
| I0408 02:20:08.237682 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.148936 | |
| I0408 02:20:08.237694 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.761364 | |
| I0408 02:20:08.237707 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.234043 | |
| I0408 02:20:08.237722 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 4.10973 (* 1 = 4.10973 loss) | |
| I0408 02:20:08.237737 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.1961 (* 1 = 1.1961 loss) | |
| I0408 02:20:08.237751 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:20:08.237762 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00041101 | |
| I0408 02:20:08.237778 3443 sgd_solver.cpp:106] Iteration 13000, lr = 0.00981429 | |
| I0408 02:25:42.615715 3443 solver.cpp:229] Iteration 13500, loss = 6.41417 | |
| I0408 02:25:42.615835 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0888889 | |
| I0408 02:25:42.615855 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 02:25:42.615869 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.222222 | |
| I0408 02:25:42.615886 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.17528 (* 0.3 = 0.952583 loss) | |
| I0408 02:25:42.615902 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.935502 (* 0.3 = 0.28065 loss) | |
| I0408 02:25:42.615914 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0888889 | |
| I0408 02:25:42.615931 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.755682 | |
| I0408 02:25:42.615943 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.177778 | |
| I0408 02:25:42.615958 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.2288 (* 0.3 = 0.968639 loss) | |
| I0408 02:25:42.615973 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.948312 (* 0.3 = 0.284494 loss) | |
| I0408 02:25:42.615986 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.177778 | |
| I0408 02:25:42.615998 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 02:25:42.616010 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.311111 | |
| I0408 02:25:42.616025 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.01037 (* 1 = 3.01037 loss) | |
| I0408 02:25:42.616039 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.865301 (* 1 = 0.865301 loss) | |
| I0408 02:25:42.616052 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:25:42.616065 3443 solver.cpp:245] Train net output #16: total_confidence = 5.19581e-05 | |
| I0408 02:25:42.616080 3443 sgd_solver.cpp:106] Iteration 13500, lr = 0.00980714 | |
| I0408 02:31:16.686941 3443 solver.cpp:229] Iteration 14000, loss = 6.40078 | |
| I0408 02:31:16.687052 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.121951 | |
| I0408 02:31:16.687070 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 02:31:16.687084 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.317073 | |
| I0408 02:31:16.687101 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.48454 (* 0.3 = 1.04536 loss) | |
| I0408 02:31:16.687116 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.94436 (* 0.3 = 0.283308 loss) | |
| I0408 02:31:16.687129 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.146341 | |
| I0408 02:31:16.687142 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 02:31:16.687155 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.243902 | |
| I0408 02:31:16.687168 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.47854 (* 0.3 = 1.04356 loss) | |
| I0408 02:31:16.687183 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.946943 (* 0.3 = 0.284083 loss) | |
| I0408 02:31:16.687196 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.146341 | |
| I0408 02:31:16.687208 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.784091 | |
| I0408 02:31:16.687221 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.268293 | |
| I0408 02:31:16.687234 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.24218 (* 1 = 3.24218 loss) | |
| I0408 02:31:16.687249 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.87597 (* 1 = 0.87597 loss) | |
| I0408 02:31:16.687261 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:31:16.687273 3443 solver.cpp:245] Train net output #16: total_confidence = 2.6492e-06 | |
| I0408 02:31:16.687288 3443 sgd_solver.cpp:106] Iteration 14000, lr = 0.0098 | |
| I0408 02:36:50.599452 3443 solver.cpp:229] Iteration 14500, loss = 6.38503 | |
| I0408 02:36:50.599598 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.12963 | |
| I0408 02:36:50.599619 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 02:36:50.599633 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.296296 | |
| I0408 02:36:50.599661 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.11398 (* 0.3 = 0.934194 loss) | |
| I0408 02:36:50.599686 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.996034 (* 0.3 = 0.29881 loss) | |
| I0408 02:36:50.599700 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0555556 | |
| I0408 02:36:50.599714 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.704545 | |
| I0408 02:36:50.599735 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.259259 | |
| I0408 02:36:50.599750 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.16719 (* 0.3 = 0.950157 loss) | |
| I0408 02:36:50.599764 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.03688 (* 0.3 = 0.311064 loss) | |
| I0408 02:36:50.599777 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.0925926 | |
| I0408 02:36:50.599792 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.721591 | |
| I0408 02:36:50.599804 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.240741 | |
| I0408 02:36:50.599819 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.0275 (* 1 = 3.0275 loss) | |
| I0408 02:36:50.599834 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.975328 (* 1 = 0.975328 loss) | |
| I0408 02:36:50.599846 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:36:50.599859 3443 solver.cpp:245] Train net output #16: total_confidence = 2.07313e-06 | |
| I0408 02:36:50.599874 3443 sgd_solver.cpp:106] Iteration 14500, lr = 0.00979286 | |
| I0408 02:42:24.104883 3443 solver.cpp:338] Iteration 15000, Testing net (#0) | |
| I0408 02:43:05.553393 3443 solver.cpp:393] Test loss: 5.6876 | |
| I0408 02:43:05.553534 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.111799 | |
| I0408 02:43:05.553555 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.778954 | |
| I0408 02:43:05.553568 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.339578 | |
| I0408 02:43:05.553586 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.88876 (* 0.3 = 0.866627 loss) | |
| I0408 02:43:05.553601 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.730929 (* 0.3 = 0.219279 loss) | |
| I0408 02:43:05.553614 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.102841 | |
| I0408 02:43:05.553627 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.779272 | |
| I0408 02:43:05.553637 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.329391 | |
| I0408 02:43:05.553652 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 2.89455 (* 0.3 = 0.868366 loss) | |
| I0408 02:43:05.553665 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.731875 (* 0.3 = 0.219563 loss) | |
| I0408 02:43:05.553678 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.120547 | |
| I0408 02:43:05.553690 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.778136 | |
| I0408 02:43:05.553702 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.360552 | |
| I0408 02:43:05.553717 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 2.78008 (* 1 = 2.78008 loss) | |
| I0408 02:43:05.553731 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.733689 (* 1 = 0.733689 loss) | |
| I0408 02:43:05.553743 3443 solver.cpp:406] Test net output #15: total_accuracy = 0 | |
| I0408 02:43:05.553755 3443 solver.cpp:406] Test net output #16: total_confidence = 0.000918233 | |
| I0408 02:43:05.929527 3443 solver.cpp:229] Iteration 15000, loss = 6.33632 | |
| I0408 02:43:05.929599 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0869565 | |
| I0408 02:43:05.929617 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 02:43:05.929631 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.347826 | |
| I0408 02:43:05.929648 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.81354 (* 0.3 = 0.844062 loss) | |
| I0408 02:43:05.929663 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.830055 (* 0.3 = 0.249016 loss) | |
| I0408 02:43:05.929677 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0652174 | |
| I0408 02:43:05.929689 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.744318 | |
| I0408 02:43:05.929707 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.347826 | |
| I0408 02:43:05.929723 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.7482 (* 0.3 = 0.82446 loss) | |
| I0408 02:43:05.929738 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.787455 (* 0.3 = 0.236236 loss) | |
| I0408 02:43:05.929751 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.217391 | |
| I0408 02:43:05.929764 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.789773 | |
| I0408 02:43:05.929776 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.521739 | |
| I0408 02:43:05.929791 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.40939 (* 1 = 2.40939 loss) | |
| I0408 02:43:05.929805 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.720687 (* 1 = 0.720687 loss) | |
| I0408 02:43:05.929818 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:43:05.929831 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000688183 | |
| I0408 02:43:05.929846 3443 sgd_solver.cpp:106] Iteration 15000, lr = 0.00978571 | |
| I0408 02:48:39.182595 3443 solver.cpp:229] Iteration 15500, loss = 6.29518 | |
| I0408 02:48:39.182742 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.145833 | |
| I0408 02:48:39.182765 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 02:48:39.182778 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.354167 | |
| I0408 02:48:39.182796 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.88797 (* 0.3 = 0.866392 loss) | |
| I0408 02:48:39.182811 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.888101 (* 0.3 = 0.26643 loss) | |
| I0408 02:48:39.182823 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.145833 | |
| I0408 02:48:39.182837 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 02:48:39.182848 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.333333 | |
| I0408 02:48:39.182863 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.83267 (* 0.3 = 0.849802 loss) | |
| I0408 02:48:39.182878 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.852831 (* 0.3 = 0.255849 loss) | |
| I0408 02:48:39.182891 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.1875 | |
| I0408 02:48:39.182904 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 02:48:39.182916 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.354167 | |
| I0408 02:48:39.182934 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.63761 (* 1 = 2.63761 loss) | |
| I0408 02:48:39.182948 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.782233 (* 1 = 0.782233 loss) | |
| I0408 02:48:39.182961 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:48:39.182973 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000685446 | |
| I0408 02:48:39.182988 3443 sgd_solver.cpp:106] Iteration 15500, lr = 0.00977857 | |
| I0408 02:54:12.558722 3443 solver.cpp:229] Iteration 16000, loss = 6.23655 | |
| I0408 02:54:12.558851 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0961538 | |
| I0408 02:54:12.558871 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 02:54:12.558886 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.25 | |
| I0408 02:54:12.558902 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.14522 (* 0.3 = 0.943566 loss) | |
| I0408 02:54:12.558919 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.972384 (* 0.3 = 0.291715 loss) | |
| I0408 02:54:12.558933 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0576923 | |
| I0408 02:54:12.558946 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.721591 | |
| I0408 02:54:12.558959 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.269231 | |
| I0408 02:54:12.558974 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.98626 (* 0.3 = 0.895877 loss) | |
| I0408 02:54:12.558989 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.933076 (* 0.3 = 0.279923 loss) | |
| I0408 02:54:12.559001 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.115385 | |
| I0408 02:54:12.559015 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.738636 | |
| I0408 02:54:12.559026 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.307692 | |
| I0408 02:54:12.559041 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.8921 (* 1 = 2.8921 loss) | |
| I0408 02:54:12.559056 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.895257 (* 1 = 0.895257 loss) | |
| I0408 02:54:12.559068 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:54:12.559080 3443 solver.cpp:245] Train net output #16: total_confidence = 9.86625e-05 | |
| I0408 02:54:12.559094 3443 sgd_solver.cpp:106] Iteration 16000, lr = 0.00977143 | |
| I0408 02:59:45.953665 3443 solver.cpp:229] Iteration 16500, loss = 6.18739 | |
| I0408 02:59:45.953778 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0833333 | |
| I0408 02:59:45.953799 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 02:59:45.953812 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.166667 | |
| I0408 02:59:45.953830 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.59447 (* 0.3 = 1.07834 loss) | |
| I0408 02:59:45.953845 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.0439 (* 0.3 = 0.313169 loss) | |
| I0408 02:59:45.953857 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.104167 | |
| I0408 02:59:45.953871 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.75 | |
| I0408 02:59:45.953883 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.229167 | |
| I0408 02:59:45.953897 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.44302 (* 0.3 = 1.0329 loss) | |
| I0408 02:59:45.953912 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.0028 (* 0.3 = 0.30084 loss) | |
| I0408 02:59:45.953924 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.145833 | |
| I0408 02:59:45.953938 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.755682 | |
| I0408 02:59:45.953949 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.270833 | |
| I0408 02:59:45.953964 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.06971 (* 1 = 3.06971 loss) | |
| I0408 02:59:45.953979 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.931412 (* 1 = 0.931412 loss) | |
| I0408 02:59:45.953990 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 02:59:45.954004 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000232244 | |
| I0408 02:59:45.954017 3443 sgd_solver.cpp:106] Iteration 16500, lr = 0.00976429 | |
| I0408 03:05:19.326282 3443 solver.cpp:229] Iteration 17000, loss = 6.22728 | |
| I0408 03:05:19.326357 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.204545 | |
| I0408 03:05:19.326376 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 03:05:19.326390 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.318182 | |
| I0408 03:05:19.326406 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.24279 (* 0.3 = 0.972838 loss) | |
| I0408 03:05:19.326422 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.878274 (* 0.3 = 0.263482 loss) | |
| I0408 03:05:19.326436 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.136364 | |
| I0408 03:05:19.326448 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 03:05:19.326460 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.318182 | |
| I0408 03:05:19.326474 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.10526 (* 0.3 = 0.931579 loss) | |
| I0408 03:05:19.326489 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.853778 (* 0.3 = 0.256134 loss) | |
| I0408 03:05:19.326501 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.25 | |
| I0408 03:05:19.326514 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.789773 | |
| I0408 03:05:19.326526 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.409091 | |
| I0408 03:05:19.326541 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.97708 (* 1 = 2.97708 loss) | |
| I0408 03:05:19.326560 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.838919 (* 1 = 0.838919 loss) | |
| I0408 03:05:19.326571 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:05:19.326584 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000253977 | |
| I0408 03:05:19.326598 3443 sgd_solver.cpp:106] Iteration 17000, lr = 0.00975714 | |
| I0408 03:10:52.727918 3443 solver.cpp:229] Iteration 17500, loss = 6.154 | |
| I0408 03:10:52.728051 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0689655 | |
| I0408 03:10:52.728072 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.681818 | |
| I0408 03:10:52.728086 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.206897 | |
| I0408 03:10:52.728104 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.50125 (* 0.3 = 1.05038 loss) | |
| I0408 03:10:52.728119 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.2217 (* 0.3 = 0.36651 loss) | |
| I0408 03:10:52.728132 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0517241 | |
| I0408 03:10:52.728145 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.676136 | |
| I0408 03:10:52.728157 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.172414 | |
| I0408 03:10:52.728171 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.40802 (* 0.3 = 1.02241 loss) | |
| I0408 03:10:52.728186 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.19211 (* 0.3 = 0.357633 loss) | |
| I0408 03:10:52.728199 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.103448 | |
| I0408 03:10:52.728211 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.693182 | |
| I0408 03:10:52.728224 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.275862 | |
| I0408 03:10:52.728238 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.65851 (* 1 = 3.65851 loss) | |
| I0408 03:10:52.728252 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 1.26509 (* 1 = 1.26509 loss) | |
| I0408 03:10:52.728265 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:10:52.728276 3443 solver.cpp:245] Train net output #16: total_confidence = 9.95308e-06 | |
| I0408 03:10:52.728291 3443 sgd_solver.cpp:106] Iteration 17500, lr = 0.00975 | |
| I0408 03:16:26.094965 3443 solver.cpp:229] Iteration 18000, loss = 6.11283 | |
| I0408 03:16:26.095083 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.097561 | |
| I0408 03:16:26.095104 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 03:16:26.095118 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.268293 | |
| I0408 03:16:26.095135 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.06155 (* 0.3 = 0.918464 loss) | |
| I0408 03:16:26.095150 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.883058 (* 0.3 = 0.264917 loss) | |
| I0408 03:16:26.095163 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.121951 | |
| I0408 03:16:26.095176 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 03:16:26.095188 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.317073 | |
| I0408 03:16:26.095203 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.9982 (* 0.3 = 0.899461 loss) | |
| I0408 03:16:26.095218 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.848543 (* 0.3 = 0.254563 loss) | |
| I0408 03:16:26.095230 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.219512 | |
| I0408 03:16:26.095243 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.784091 | |
| I0408 03:16:26.095257 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.512195 | |
| I0408 03:16:26.095270 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.69777 (* 1 = 2.69777 loss) | |
| I0408 03:16:26.095285 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.786169 (* 1 = 0.786169 loss) | |
| I0408 03:16:26.095299 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:16:26.095335 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000815731 | |
| I0408 03:16:26.095352 3443 sgd_solver.cpp:106] Iteration 18000, lr = 0.00974286 | |
| I0408 03:21:59.491999 3443 solver.cpp:229] Iteration 18500, loss = 6.05918 | |
| I0408 03:21:59.492120 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.162791 | |
| I0408 03:21:59.492139 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 03:21:59.492153 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.325581 | |
| I0408 03:21:59.492169 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.96062 (* 0.3 = 0.888186 loss) | |
| I0408 03:21:59.492184 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.806971 (* 0.3 = 0.242091 loss) | |
| I0408 03:21:59.492197 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.186047 | |
| I0408 03:21:59.492210 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 03:21:59.492223 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.302326 | |
| I0408 03:21:59.492238 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.73543 (* 0.3 = 0.82063 loss) | |
| I0408 03:21:59.492252 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.757007 (* 0.3 = 0.227102 loss) | |
| I0408 03:21:59.492264 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.162791 | |
| I0408 03:21:59.492277 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.795455 | |
| I0408 03:21:59.492290 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.348837 | |
| I0408 03:21:59.492305 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.76149 (* 1 = 2.76149 loss) | |
| I0408 03:21:59.492318 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.722848 (* 1 = 0.722848 loss) | |
| I0408 03:21:59.492331 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:21:59.492342 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000268515 | |
| I0408 03:21:59.492357 3443 sgd_solver.cpp:106] Iteration 18500, lr = 0.00973571 | |
| I0408 03:27:32.865396 3443 solver.cpp:229] Iteration 19000, loss = 6.07331 | |
| I0408 03:27:32.865542 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.125 | |
| I0408 03:27:32.865563 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 03:27:32.865577 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.395833 | |
| I0408 03:27:32.865593 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.93278 (* 0.3 = 0.879833 loss) | |
| I0408 03:27:32.865609 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.867772 (* 0.3 = 0.260332 loss) | |
| I0408 03:27:32.865622 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0833333 | |
| I0408 03:27:32.865635 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.744318 | |
| I0408 03:27:32.865648 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.333333 | |
| I0408 03:27:32.865661 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.13244 (* 0.3 = 0.939732 loss) | |
| I0408 03:27:32.865676 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.920409 (* 0.3 = 0.276123 loss) | |
| I0408 03:27:32.865689 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.166667 | |
| I0408 03:27:32.865701 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 03:27:32.865713 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.354167 | |
| I0408 03:27:32.865727 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.89764 (* 1 = 2.89764 loss) | |
| I0408 03:27:32.865741 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.823988 (* 1 = 0.823988 loss) | |
| I0408 03:27:32.865754 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:27:32.865767 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000703157 | |
| I0408 03:27:32.865782 3443 sgd_solver.cpp:106] Iteration 19000, lr = 0.00972857 | |
| I0408 03:33:06.261752 3443 solver.cpp:229] Iteration 19500, loss = 6.02707 | |
| I0408 03:33:06.261965 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.139535 | |
| I0408 03:33:06.261996 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 03:33:06.262018 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.302326 | |
| I0408 03:33:06.262045 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.86896 (* 0.3 = 0.860689 loss) | |
| I0408 03:33:06.262069 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.841048 (* 0.3 = 0.252314 loss) | |
| I0408 03:33:06.262090 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.139535 | |
| I0408 03:33:06.262110 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.767045 | |
| I0408 03:33:06.262131 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.302326 | |
| I0408 03:33:06.262157 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.82822 (* 0.3 = 0.848467 loss) | |
| I0408 03:33:06.262186 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.843234 (* 0.3 = 0.25297 loss) | |
| I0408 03:33:06.262208 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.209302 | |
| I0408 03:33:06.262230 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.789773 | |
| I0408 03:33:06.262251 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.372093 | |
| I0408 03:33:06.262277 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.59688 (* 1 = 2.59688 loss) | |
| I0408 03:33:06.262302 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.740166 (* 1 = 0.740166 loss) | |
| I0408 03:33:06.262325 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:33:06.262344 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000361148 | |
| I0408 03:33:06.262369 3443 sgd_solver.cpp:106] Iteration 19500, lr = 0.00972143 | |
| I0408 03:38:39.419463 3443 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_20000.caffemodel | |
| I0408 03:38:40.408098 3443 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_20000.solverstate | |
| I0408 03:38:40.663452 3443 solver.cpp:338] Iteration 20000, Testing net (#0) | |
| I0408 03:39:22.200975 3443 solver.cpp:393] Test loss: 5.47749 | |
| I0408 03:39:22.201160 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.121362 | |
| I0408 03:39:22.201181 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.782772 | |
| I0408 03:39:22.201195 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.312271 | |
| I0408 03:39:22.201212 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 3.10695 (* 0.3 = 0.932086 loss) | |
| I0408 03:39:22.201227 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.80271 (* 0.3 = 0.240813 loss) | |
| I0408 03:39:22.201241 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.187429 | |
| I0408 03:39:22.201252 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.791636 | |
| I0408 03:39:22.201264 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.422584 | |
| I0408 03:39:22.201278 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 2.80781 (* 0.3 = 0.842343 loss) | |
| I0408 03:39:22.201292 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.768699 (* 0.3 = 0.23061 loss) | |
| I0408 03:39:22.201304 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.248012 | |
| I0408 03:39:22.201318 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.805499 | |
| I0408 03:39:22.201328 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.517742 | |
| I0408 03:39:22.201341 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 2.53828 (* 1 = 2.53828 loss) | |
| I0408 03:39:22.201355 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.693356 (* 1 = 0.693356 loss) | |
| I0408 03:39:22.201367 3443 solver.cpp:406] Test net output #15: total_accuracy = 0 | |
| I0408 03:39:22.201378 3443 solver.cpp:406] Test net output #16: total_confidence = 0.00353288 | |
| I0408 03:39:22.578279 3443 solver.cpp:229] Iteration 20000, loss = 5.90747 | |
| I0408 03:39:22.578366 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0851064 | |
| I0408 03:39:22.578397 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 03:39:22.578418 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.212766 | |
| I0408 03:39:22.578445 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.17171 (* 0.3 = 0.951513 loss) | |
| I0408 03:39:22.578471 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.906176 (* 0.3 = 0.271853 loss) | |
| I0408 03:39:22.578493 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.148936 | |
| I0408 03:39:22.578516 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 03:39:22.578543 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.276596 | |
| I0408 03:39:22.578569 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.98058 (* 0.3 = 0.894174 loss) | |
| I0408 03:39:22.578594 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.867251 (* 0.3 = 0.260175 loss) | |
| I0408 03:39:22.578616 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.191489 | |
| I0408 03:39:22.578637 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.784091 | |
| I0408 03:39:22.578660 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.340426 | |
| I0408 03:39:22.578683 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.74857 (* 1 = 2.74857 loss) | |
| I0408 03:39:22.578711 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.777717 (* 1 = 0.777717 loss) | |
| I0408 03:39:22.578734 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:39:22.578755 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000365537 | |
| I0408 03:39:22.578779 3443 sgd_solver.cpp:106] Iteration 20000, lr = 0.00971429 | |
| I0408 03:44:56.047286 3443 solver.cpp:229] Iteration 20500, loss = 5.96394 | |
| I0408 03:44:56.047472 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0232558 | |
| I0408 03:44:56.047494 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 03:44:56.047508 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.209302 | |
| I0408 03:44:56.047524 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.49207 (* 0.3 = 1.04762 loss) | |
| I0408 03:44:56.047540 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.990423 (* 0.3 = 0.297127 loss) | |
| I0408 03:44:56.047554 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0697674 | |
| I0408 03:44:56.047567 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.767045 | |
| I0408 03:44:56.047590 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.209302 | |
| I0408 03:44:56.047605 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.37157 (* 0.3 = 1.01147 loss) | |
| I0408 03:44:56.047621 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.895009 (* 0.3 = 0.268503 loss) | |
| I0408 03:44:56.047632 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.209302 | |
| I0408 03:44:56.047646 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.784091 | |
| I0408 03:44:56.047658 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.302326 | |
| I0408 03:44:56.047673 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.95396 (* 1 = 2.95396 loss) | |
| I0408 03:44:56.047688 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.840022 (* 1 = 0.840022 loss) | |
| I0408 03:44:56.047700 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:44:56.047713 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00479399 | |
| I0408 03:44:56.047727 3443 sgd_solver.cpp:106] Iteration 20500, lr = 0.00970714 | |
| I0408 03:50:29.756602 3443 solver.cpp:229] Iteration 21000, loss = 5.89677 | |
| I0408 03:50:29.756731 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170213 | |
| I0408 03:50:29.756750 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 03:50:29.756764 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.361702 | |
| I0408 03:50:29.756781 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.1291 (* 0.3 = 0.938729 loss) | |
| I0408 03:50:29.756796 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.919596 (* 0.3 = 0.275879 loss) | |
| I0408 03:50:29.756809 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.106383 | |
| I0408 03:50:29.756824 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 03:50:29.756835 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.255319 | |
| I0408 03:50:29.756850 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.15096 (* 0.3 = 0.945289 loss) | |
| I0408 03:50:29.756865 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.896236 (* 0.3 = 0.268871 loss) | |
| I0408 03:50:29.756876 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.148936 | |
| I0408 03:50:29.756888 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.767045 | |
| I0408 03:50:29.756901 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.404255 | |
| I0408 03:50:29.756916 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.88524 (* 1 = 2.88524 loss) | |
| I0408 03:50:29.756932 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.827257 (* 1 = 0.827257 loss) | |
| I0408 03:50:29.756944 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:50:29.756958 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000214557 | |
| I0408 03:50:29.756973 3443 sgd_solver.cpp:106] Iteration 21000, lr = 0.0097 | |
| I0408 03:56:03.296778 3443 solver.cpp:229] Iteration 21500, loss = 5.8648 | |
| I0408 03:56:03.296967 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.102041 | |
| I0408 03:56:03.296989 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 03:56:03.297003 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.326531 | |
| I0408 03:56:03.297021 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.91291 (* 0.3 = 0.873874 loss) | |
| I0408 03:56:03.297037 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.847433 (* 0.3 = 0.25423 loss) | |
| I0408 03:56:03.297050 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0408163 | |
| I0408 03:56:03.297063 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.732955 | |
| I0408 03:56:03.297076 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.346939 | |
| I0408 03:56:03.297091 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.90866 (* 0.3 = 0.872597 loss) | |
| I0408 03:56:03.297106 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.840847 (* 0.3 = 0.252254 loss) | |
| I0408 03:56:03.297117 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.183673 | |
| I0408 03:56:03.297130 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 03:56:03.297144 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.469388 | |
| I0408 03:56:03.297159 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.58836 (* 1 = 2.58836 loss) | |
| I0408 03:56:03.297173 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.755498 (* 1 = 0.755498 loss) | |
| I0408 03:56:03.297186 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 03:56:03.297199 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000526112 | |
| I0408 03:56:03.297214 3443 sgd_solver.cpp:106] Iteration 21500, lr = 0.00969286 | |
| I0408 04:01:37.219444 3443 solver.cpp:229] Iteration 22000, loss = 5.84008 | |
| I0408 04:01:37.219607 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.121212 | |
| I0408 04:01:37.219630 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 04:01:37.219643 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.424242 | |
| I0408 04:01:37.219660 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.32698 (* 0.3 = 0.998095 loss) | |
| I0408 04:01:37.219676 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.891549 (* 0.3 = 0.267465 loss) | |
| I0408 04:01:37.219696 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.242424 | |
| I0408 04:01:37.219722 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 04:01:37.219746 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.393939 | |
| I0408 04:01:37.219763 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.26208 (* 0.3 = 0.978625 loss) | |
| I0408 04:01:37.219777 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.878891 (* 0.3 = 0.263667 loss) | |
| I0408 04:01:37.219790 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.181818 | |
| I0408 04:01:37.219804 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.8125 | |
| I0408 04:01:37.219815 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.424242 | |
| I0408 04:01:37.219830 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 3.23786 (* 1 = 3.23786 loss) | |
| I0408 04:01:37.219846 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.77744 (* 1 = 0.77744 loss) | |
| I0408 04:01:37.219858 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:01:37.219871 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000834102 | |
| I0408 04:01:37.219887 3443 sgd_solver.cpp:106] Iteration 22000, lr = 0.00968571 | |
| I0408 04:07:10.587605 3443 solver.cpp:229] Iteration 22500, loss = 5.79453 | |
| I0408 04:07:10.587749 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0909091 | |
| I0408 04:07:10.587769 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 04:07:10.587782 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.431818 | |
| I0408 04:07:10.587800 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.7063 (* 0.3 = 0.81189 loss) | |
| I0408 04:07:10.587815 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.761748 (* 0.3 = 0.228524 loss) | |
| I0408 04:07:10.587828 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.136364 | |
| I0408 04:07:10.587841 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 04:07:10.587853 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.477273 | |
| I0408 04:07:10.587867 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.65143 (* 0.3 = 0.79543 loss) | |
| I0408 04:07:10.587882 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.724082 (* 0.3 = 0.217225 loss) | |
| I0408 04:07:10.587894 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.204545 | |
| I0408 04:07:10.587908 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.789773 | |
| I0408 04:07:10.587929 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.477273 | |
| I0408 04:07:10.587957 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.49052 (* 1 = 2.49052 loss) | |
| I0408 04:07:10.587975 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.664172 (* 1 = 0.664172 loss) | |
| I0408 04:07:10.587990 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:07:10.588001 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00108864 | |
| I0408 04:07:10.588016 3443 sgd_solver.cpp:106] Iteration 22500, lr = 0.00967857 | |
| I0408 04:12:43.973495 3443 solver.cpp:229] Iteration 23000, loss = 5.75905 | |
| I0408 04:12:43.973620 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0769231 | |
| I0408 04:12:43.973641 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.727273 | |
| I0408 04:12:43.973655 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.269231 | |
| I0408 04:12:43.973672 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.21488 (* 0.3 = 0.964463 loss) | |
| I0408 04:12:43.973687 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.987613 (* 0.3 = 0.296284 loss) | |
| I0408 04:12:43.973701 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.0961538 | |
| I0408 04:12:43.973713 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.732955 | |
| I0408 04:12:43.973726 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.365385 | |
| I0408 04:12:43.973740 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.20837 (* 0.3 = 0.962512 loss) | |
| I0408 04:12:43.973754 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.987533 (* 0.3 = 0.29626 loss) | |
| I0408 04:12:43.973767 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.134615 | |
| I0408 04:12:43.973779 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.738636 | |
| I0408 04:12:43.973793 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.346154 | |
| I0408 04:12:43.973806 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.90652 (* 1 = 2.90652 loss) | |
| I0408 04:12:43.973820 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.912728 (* 1 = 0.912728 loss) | |
| I0408 04:12:43.973832 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:12:43.973845 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00286899 | |
| I0408 04:12:43.973860 3443 sgd_solver.cpp:106] Iteration 23000, lr = 0.00967143 | |
| I0408 04:18:17.359911 3443 solver.cpp:229] Iteration 23500, loss = 5.72152 | |
| I0408 04:18:17.360080 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0697674 | |
| I0408 04:18:17.360102 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 04:18:17.360116 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.372093 | |
| I0408 04:18:17.360132 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.88174 (* 0.3 = 0.864522 loss) | |
| I0408 04:18:17.360148 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.76453 (* 0.3 = 0.229359 loss) | |
| I0408 04:18:17.360160 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.255814 | |
| I0408 04:18:17.360173 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 04:18:17.360186 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.465116 | |
| I0408 04:18:17.360200 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.7252 (* 0.3 = 0.817559 loss) | |
| I0408 04:18:17.360215 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.739986 (* 0.3 = 0.221996 loss) | |
| I0408 04:18:17.360229 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.255814 | |
| I0408 04:18:17.360240 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.801136 | |
| I0408 04:18:17.360252 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.488372 | |
| I0408 04:18:17.360267 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.45667 (* 1 = 2.45667 loss) | |
| I0408 04:18:17.360281 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.673545 (* 1 = 0.673545 loss) | |
| I0408 04:18:17.360294 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:18:17.360306 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000185179 | |
| I0408 04:18:17.360321 3443 sgd_solver.cpp:106] Iteration 23500, lr = 0.00966429 | |
| I0408 04:23:51.076306 3443 solver.cpp:229] Iteration 24000, loss = 5.6634 | |
| I0408 04:23:51.076447 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.130435 | |
| I0408 04:23:51.076468 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 04:23:51.076493 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.391304 | |
| I0408 04:23:51.076520 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.92501 (* 0.3 = 0.877504 loss) | |
| I0408 04:23:51.076537 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.841139 (* 0.3 = 0.252342 loss) | |
| I0408 04:23:51.076550 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.152174 | |
| I0408 04:23:51.076563 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 04:23:51.076575 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.391304 | |
| I0408 04:23:51.076591 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.79658 (* 0.3 = 0.838975 loss) | |
| I0408 04:23:51.076604 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.832479 (* 0.3 = 0.249744 loss) | |
| I0408 04:23:51.076617 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.217391 | |
| I0408 04:23:51.076629 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.778409 | |
| I0408 04:23:51.076642 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.5 | |
| I0408 04:23:51.076656 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.48546 (* 1 = 2.48546 loss) | |
| I0408 04:23:51.076670 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.723716 (* 1 = 0.723716 loss) | |
| I0408 04:23:51.076683 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:23:51.076694 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00111612 | |
| I0408 04:23:51.076709 3443 sgd_solver.cpp:106] Iteration 24000, lr = 0.00965714 | |
| I0408 04:29:25.118901 3443 solver.cpp:229] Iteration 24500, loss = 5.68391 | |
| I0408 04:29:25.119024 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.102041 | |
| I0408 04:29:25.119053 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 04:29:25.119077 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.387755 | |
| I0408 04:29:25.119105 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.69484 (* 0.3 = 0.808452 loss) | |
| I0408 04:29:25.119134 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.854493 (* 0.3 = 0.256348 loss) | |
| I0408 04:29:25.119161 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.142857 | |
| I0408 04:29:25.119185 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0408 04:29:25.119205 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.489796 | |
| I0408 04:29:25.119233 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.59141 (* 0.3 = 0.777422 loss) | |
| I0408 04:29:25.119261 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.839809 (* 0.3 = 0.251943 loss) | |
| I0408 04:29:25.119282 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.22449 | |
| I0408 04:29:25.119304 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.778409 | |
| I0408 04:29:25.119348 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.612245 | |
| I0408 04:29:25.119376 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.26674 (* 1 = 2.26674 loss) | |
| I0408 04:29:25.119403 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.700507 (* 1 = 0.700507 loss) | |
| I0408 04:29:25.119426 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:29:25.119449 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000385832 | |
| I0408 04:29:25.119474 3443 sgd_solver.cpp:106] Iteration 24500, lr = 0.00965 | |
| I0408 04:34:58.115461 3443 solver.cpp:338] Iteration 25000, Testing net (#0) | |
| I0408 04:35:39.420289 3443 solver.cpp:393] Test loss: 4.91618 | |
| I0408 04:35:39.420397 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.137866 | |
| I0408 04:35:39.420418 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.786045 | |
| I0408 04:35:39.420434 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.386392 | |
| I0408 04:35:39.420454 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.87346 (* 0.3 = 0.862037 loss) | |
| I0408 04:35:39.420470 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.73311 (* 0.3 = 0.219933 loss) | |
| I0408 04:35:39.420483 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.239812 | |
| I0408 04:35:39.420496 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.808728 | |
| I0408 04:35:39.420508 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.507788 | |
| I0408 04:35:39.420522 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 2.51268 (* 0.3 = 0.753803 loss) | |
| I0408 04:35:39.420537 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.656623 (* 0.3 = 0.196987 loss) | |
| I0408 04:35:39.420549 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.301843 | |
| I0408 04:35:39.420562 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.822955 | |
| I0408 04:35:39.420572 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.609965 | |
| I0408 04:35:39.420586 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 2.28575 (* 1 = 2.28575 loss) | |
| I0408 04:35:39.420601 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.597672 (* 1 = 0.597672 loss) | |
| I0408 04:35:39.420614 3443 solver.cpp:406] Test net output #15: total_accuracy = 0 | |
| I0408 04:35:39.420625 3443 solver.cpp:406] Test net output #16: total_confidence = 0.0125953 | |
| I0408 04:35:39.799145 3443 solver.cpp:229] Iteration 25000, loss = 5.61565 | |
| I0408 04:35:39.799216 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.181818 | |
| I0408 04:35:39.799234 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 04:35:39.799247 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.386364 | |
| I0408 04:35:39.799264 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.99888 (* 0.3 = 0.899664 loss) | |
| I0408 04:35:39.799280 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.819139 (* 0.3 = 0.245742 loss) | |
| I0408 04:35:39.799293 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.204545 | |
| I0408 04:35:39.799306 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 04:35:39.799334 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.454545 | |
| I0408 04:35:39.799351 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.68279 (* 0.3 = 0.804837 loss) | |
| I0408 04:35:39.799366 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.741027 (* 0.3 = 0.222308 loss) | |
| I0408 04:35:39.799379 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.25 | |
| I0408 04:35:39.799392 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.801136 | |
| I0408 04:35:39.799406 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.5 | |
| I0408 04:35:39.799419 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.52897 (* 1 = 2.52897 loss) | |
| I0408 04:35:39.799434 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.689765 (* 1 = 0.689765 loss) | |
| I0408 04:35:39.799446 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:35:39.799459 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00927546 | |
| I0408 04:35:39.799474 3443 sgd_solver.cpp:106] Iteration 25000, lr = 0.00964286 | |
| I0408 04:41:13.217890 3443 solver.cpp:229] Iteration 25500, loss = 5.57057 | |
| I0408 04:41:13.218024 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1 | |
| I0408 04:41:13.218044 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 04:41:13.218057 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.28 | |
| I0408 04:41:13.218073 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.05437 (* 0.3 = 0.916312 loss) | |
| I0408 04:41:13.218088 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.909091 (* 0.3 = 0.272727 loss) | |
| I0408 04:41:13.218101 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.08 | |
| I0408 04:41:13.218114 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0408 04:41:13.218127 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.36 | |
| I0408 04:41:13.218140 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.89131 (* 0.3 = 0.867392 loss) | |
| I0408 04:41:13.218155 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.863423 (* 0.3 = 0.259027 loss) | |
| I0408 04:41:13.218168 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.16 | |
| I0408 04:41:13.218180 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.755682 | |
| I0408 04:41:13.218192 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.42 | |
| I0408 04:41:13.218207 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.55978 (* 1 = 2.55978 loss) | |
| I0408 04:41:13.218221 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.768538 (* 1 = 0.768538 loss) | |
| I0408 04:41:13.218233 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:41:13.218245 3443 solver.cpp:245] Train net output #16: total_confidence = 0.000503808 | |
| I0408 04:41:13.218260 3443 sgd_solver.cpp:106] Iteration 25500, lr = 0.00963571 | |
| I0408 04:46:46.612912 3443 solver.cpp:229] Iteration 26000, loss = 5.57131 | |
| I0408 04:46:46.613049 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1 | |
| I0408 04:46:46.613070 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 04:46:46.613083 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.28 | |
| I0408 04:46:46.613101 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.07281 (* 0.3 = 0.921844 loss) | |
| I0408 04:46:46.613116 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.964706 (* 0.3 = 0.289412 loss) | |
| I0408 04:46:46.613128 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.16 | |
| I0408 04:46:46.613142 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.744318 | |
| I0408 04:46:46.613154 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.46 | |
| I0408 04:46:46.613168 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.86093 (* 0.3 = 0.858279 loss) | |
| I0408 04:46:46.613183 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.934049 (* 0.3 = 0.280215 loss) | |
| I0408 04:46:46.613195 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.28 | |
| I0408 04:46:46.613207 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.784091 | |
| I0408 04:46:46.613219 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.52 | |
| I0408 04:46:46.613234 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.6548 (* 1 = 2.6548 loss) | |
| I0408 04:46:46.613248 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.828316 (* 1 = 0.828316 loss) | |
| I0408 04:46:46.613261 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:46:46.613272 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00319384 | |
| I0408 04:46:46.613287 3443 sgd_solver.cpp:106] Iteration 26000, lr = 0.00962857 | |
| I0408 04:52:19.982720 3443 solver.cpp:229] Iteration 26500, loss = 5.45498 | |
| I0408 04:52:19.982820 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.159091 | |
| I0408 04:52:19.982839 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 04:52:19.982853 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.295455 | |
| I0408 04:52:19.982870 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.03894 (* 0.3 = 0.911683 loss) | |
| I0408 04:52:19.982885 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.84789 (* 0.3 = 0.254367 loss) | |
| I0408 04:52:19.982897 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.204545 | |
| I0408 04:52:19.982910 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 04:52:19.982923 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.454545 | |
| I0408 04:52:19.982936 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.7609 (* 0.3 = 0.82827 loss) | |
| I0408 04:52:19.982951 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.824445 (* 0.3 = 0.247334 loss) | |
| I0408 04:52:19.982964 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.340909 | |
| I0408 04:52:19.982976 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 04:52:19.982988 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.568182 | |
| I0408 04:52:19.983003 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.31374 (* 1 = 2.31374 loss) | |
| I0408 04:52:19.983017 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.693164 (* 1 = 0.693164 loss) | |
| I0408 04:52:19.983029 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:52:19.983042 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00202379 | |
| I0408 04:52:19.983057 3443 sgd_solver.cpp:106] Iteration 26500, lr = 0.00962143 | |
| I0408 04:57:53.373890 3443 solver.cpp:229] Iteration 27000, loss = 5.41715 | |
| I0408 04:57:53.374053 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.209302 | |
| I0408 04:57:53.374074 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 04:57:53.374089 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.325581 | |
| I0408 04:57:53.374104 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.71844 (* 0.3 = 0.815532 loss) | |
| I0408 04:57:53.374120 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.735085 (* 0.3 = 0.220525 loss) | |
| I0408 04:57:53.374133 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.27907 | |
| I0408 04:57:53.374145 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 04:57:53.374157 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.395349 | |
| I0408 04:57:53.374172 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.56507 (* 0.3 = 0.76952 loss) | |
| I0408 04:57:53.374186 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.713445 (* 0.3 = 0.214033 loss) | |
| I0408 04:57:53.374199 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.255814 | |
| I0408 04:57:53.374212 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.795455 | |
| I0408 04:57:53.374223 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.55814 | |
| I0408 04:57:53.374241 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.34383 (* 1 = 2.34383 loss) | |
| I0408 04:57:53.374258 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.655706 (* 1 = 0.655706 loss) | |
| I0408 04:57:53.374270 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 04:57:53.374284 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0113597 | |
| I0408 04:57:53.374297 3443 sgd_solver.cpp:106] Iteration 27000, lr = 0.00961429 | |
| I0408 05:03:26.762840 3443 solver.cpp:229] Iteration 27500, loss = 5.44701 | |
| I0408 05:03:26.762984 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.302326 | |
| I0408 05:03:26.763006 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0408 05:03:26.763020 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.488372 | |
| I0408 05:03:26.763036 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.60715 (* 0.3 = 0.782145 loss) | |
| I0408 05:03:26.763052 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.711584 (* 0.3 = 0.213475 loss) | |
| I0408 05:03:26.763065 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.255814 | |
| I0408 05:03:26.763078 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 05:03:26.763092 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.581395 | |
| I0408 05:03:26.763105 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.29083 (* 0.3 = 0.68725 loss) | |
| I0408 05:03:26.763120 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.632539 (* 0.3 = 0.189762 loss) | |
| I0408 05:03:26.763133 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.418605 | |
| I0408 05:03:26.763145 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.840909 | |
| I0408 05:03:26.763159 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.627907 | |
| I0408 05:03:26.763172 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.93984 (* 1 = 1.93984 loss) | |
| I0408 05:03:26.763186 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.564072 (* 1 = 0.564072 loss) | |
| I0408 05:03:26.763200 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:03:26.763211 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00527866 | |
| I0408 05:03:26.763226 3443 sgd_solver.cpp:106] Iteration 27500, lr = 0.00960714 | |
| I0408 05:09:00.148047 3443 solver.cpp:229] Iteration 28000, loss = 5.37176 | |
| I0408 05:09:00.148181 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.229167 | |
| I0408 05:09:00.148202 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 05:09:00.148216 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.354167 | |
| I0408 05:09:00.148233 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.16015 (* 0.3 = 0.948046 loss) | |
| I0408 05:09:00.148248 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.92977 (* 0.3 = 0.278931 loss) | |
| I0408 05:09:00.148262 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.25 | |
| I0408 05:09:00.148274 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 05:09:00.148295 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.333333 | |
| I0408 05:09:00.148310 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.09371 (* 0.3 = 0.928112 loss) | |
| I0408 05:09:00.148325 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.929118 (* 0.3 = 0.278736 loss) | |
| I0408 05:09:00.148337 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.270833 | |
| I0408 05:09:00.148350 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.795455 | |
| I0408 05:09:00.148363 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.479167 | |
| I0408 05:09:00.148377 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.67246 (* 1 = 2.67246 loss) | |
| I0408 05:09:00.148392 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.761312 (* 1 = 0.761312 loss) | |
| I0408 05:09:00.148404 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:09:00.148416 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00234873 | |
| I0408 05:09:00.148432 3443 sgd_solver.cpp:106] Iteration 28000, lr = 0.0096 | |
| I0408 05:14:33.524186 3443 solver.cpp:229] Iteration 28500, loss = 5.34711 | |
| I0408 05:14:33.524307 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.125 | |
| I0408 05:14:33.524334 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 05:14:33.524349 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.25 | |
| I0408 05:14:33.524365 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.23819 (* 0.3 = 0.971458 loss) | |
| I0408 05:14:33.524381 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.925843 (* 0.3 = 0.277753 loss) | |
| I0408 05:14:33.524394 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.1875 | |
| I0408 05:14:33.524406 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 05:14:33.524420 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.270833 | |
| I0408 05:14:33.524433 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.94815 (* 0.3 = 0.884445 loss) | |
| I0408 05:14:33.524447 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.861943 (* 0.3 = 0.258583 loss) | |
| I0408 05:14:33.524461 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.354167 | |
| I0408 05:14:33.524472 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.818182 | |
| I0408 05:14:33.524484 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.520833 | |
| I0408 05:14:33.524498 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.5553 (* 1 = 2.5553 loss) | |
| I0408 05:14:33.524513 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.751491 (* 1 = 0.751491 loss) | |
| I0408 05:14:33.524525 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:14:33.524538 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00175085 | |
| I0408 05:14:33.524552 3443 sgd_solver.cpp:106] Iteration 28500, lr = 0.00959286 | |
| I0408 05:20:06.928674 3443 solver.cpp:229] Iteration 29000, loss = 5.27811 | |
| I0408 05:20:06.928879 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170213 | |
| I0408 05:20:06.928901 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 05:20:06.928916 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.382979 | |
| I0408 05:20:06.928937 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.80283 (* 0.3 = 0.840849 loss) | |
| I0408 05:20:06.928953 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.838168 (* 0.3 = 0.25145 loss) | |
| I0408 05:20:06.928966 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.170213 | |
| I0408 05:20:06.928980 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 05:20:06.928992 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.425532 | |
| I0408 05:20:06.929008 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.59605 (* 0.3 = 0.778814 loss) | |
| I0408 05:20:06.929023 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.740494 (* 0.3 = 0.222148 loss) | |
| I0408 05:20:06.929035 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.382979 | |
| I0408 05:20:06.929047 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 05:20:06.929059 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.638298 | |
| I0408 05:20:06.929075 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.99669 (* 1 = 1.99669 loss) | |
| I0408 05:20:06.929090 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.595919 (* 1 = 0.595919 loss) | |
| I0408 05:20:06.929102 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:20:06.929114 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00155956 | |
| I0408 05:20:06.929131 3443 sgd_solver.cpp:106] Iteration 29000, lr = 0.00958571 | |
| I0408 05:25:40.312585 3443 solver.cpp:229] Iteration 29500, loss = 5.24723 | |
| I0408 05:25:40.312729 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.166667 | |
| I0408 05:25:40.312752 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 05:25:40.312765 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.238095 | |
| I0408 05:25:40.312783 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.93647 (* 0.3 = 0.88094 loss) | |
| I0408 05:25:40.312798 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.830184 (* 0.3 = 0.249055 loss) | |
| I0408 05:25:40.312811 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.214286 | |
| I0408 05:25:40.312824 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 05:25:40.312836 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.380952 | |
| I0408 05:25:40.312850 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.71661 (* 0.3 = 0.814982 loss) | |
| I0408 05:25:40.312865 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.750702 (* 0.3 = 0.225211 loss) | |
| I0408 05:25:40.312878 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.309524 | |
| I0408 05:25:40.312891 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.8125 | |
| I0408 05:25:40.312903 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.571429 | |
| I0408 05:25:40.312921 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.41253 (* 1 = 2.41253 loss) | |
| I0408 05:25:40.312942 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.642644 (* 1 = 0.642644 loss) | |
| I0408 05:25:40.312954 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:25:40.312966 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00615939 | |
| I0408 05:25:40.312990 3443 sgd_solver.cpp:106] Iteration 29500, lr = 0.00957857 | |
| I0408 05:31:13.343006 3443 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_30000.caffemodel | |
| I0408 05:31:13.919284 3443 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_30000.solverstate | |
| I0408 05:31:14.174617 3443 solver.cpp:338] Iteration 30000, Testing net (#0) | |
| I0408 05:31:55.663022 3443 solver.cpp:393] Test loss: 4.97754 | |
| I0408 05:31:55.663162 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.145619 | |
| I0408 05:31:55.663182 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.789954 | |
| I0408 05:31:55.663197 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.397102 | |
| I0408 05:31:55.663213 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.90851 (* 0.3 = 0.872554 loss) | |
| I0408 05:31:55.663228 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.732237 (* 0.3 = 0.219671 loss) | |
| I0408 05:31:55.663241 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.242848 | |
| I0408 05:31:55.663254 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.8115 | |
| I0408 05:31:55.663267 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.524725 | |
| I0408 05:31:55.663282 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 2.5926 (* 0.3 = 0.777781 loss) | |
| I0408 05:31:55.663296 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.659972 (* 0.3 = 0.197992 loss) | |
| I0408 05:31:55.663308 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.326873 | |
| I0408 05:31:55.663334 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.831501 | |
| I0408 05:31:55.663348 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.63838 | |
| I0408 05:31:55.663362 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 2.31849 (* 1 = 2.31849 loss) | |
| I0408 05:31:55.663377 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.591059 (* 1 = 0.591059 loss) | |
| I0408 05:31:55.663388 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.004 | |
| I0408 05:31:55.663401 3443 solver.cpp:406] Test net output #16: total_confidence = 0.0169928 | |
| I0408 05:31:56.036801 3443 solver.cpp:229] Iteration 30000, loss = 5.21011 | |
| I0408 05:31:56.036875 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.116279 | |
| I0408 05:31:56.036895 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 05:31:56.036907 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.232558 | |
| I0408 05:31:56.036924 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.25716 (* 0.3 = 0.977147 loss) | |
| I0408 05:31:56.036941 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.869196 (* 0.3 = 0.260759 loss) | |
| I0408 05:31:56.036953 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.116279 | |
| I0408 05:31:56.036967 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 05:31:56.036979 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.302326 | |
| I0408 05:31:56.036994 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.17902 (* 0.3 = 0.953705 loss) | |
| I0408 05:31:56.037009 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.829304 (* 0.3 = 0.248791 loss) | |
| I0408 05:31:56.037022 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.232558 | |
| I0408 05:31:56.037035 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.8125 | |
| I0408 05:31:56.037047 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.44186 | |
| I0408 05:31:56.037062 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.60331 (* 1 = 2.60331 loss) | |
| I0408 05:31:56.037076 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.68663 (* 1 = 0.68663 loss) | |
| I0408 05:31:56.037089 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:31:56.037104 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0220163 | |
| I0408 05:31:56.037120 3443 sgd_solver.cpp:106] Iteration 30000, lr = 0.00957143 | |
| I0408 05:37:29.441752 3443 solver.cpp:229] Iteration 30500, loss = 5.2441 | |
| I0408 05:37:29.441911 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.109091 | |
| I0408 05:37:29.441943 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.715909 | |
| I0408 05:37:29.441968 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.309091 | |
| I0408 05:37:29.441999 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.31876 (* 0.3 = 0.995627 loss) | |
| I0408 05:37:29.442025 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.07219 (* 0.3 = 0.321658 loss) | |
| I0408 05:37:29.442047 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.145455 | |
| I0408 05:37:29.442071 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.727273 | |
| I0408 05:37:29.442093 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.436364 | |
| I0408 05:37:29.442119 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.85704 (* 0.3 = 0.857112 loss) | |
| I0408 05:37:29.442144 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.931557 (* 0.3 = 0.279467 loss) | |
| I0408 05:37:29.442167 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.290909 | |
| I0408 05:37:29.442190 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.772727 | |
| I0408 05:37:29.442212 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.545455 | |
| I0408 05:37:29.442239 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.44747 (* 1 = 2.44747 loss) | |
| I0408 05:37:29.442265 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.790017 (* 1 = 0.790017 loss) | |
| I0408 05:37:29.442286 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:37:29.442306 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0100726 | |
| I0408 05:37:29.442329 3443 sgd_solver.cpp:106] Iteration 30500, lr = 0.00956429 | |
| I0408 05:43:02.813120 3443 solver.cpp:229] Iteration 31000, loss = 5.16199 | |
| I0408 05:43:02.813248 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.111111 | |
| I0408 05:43:02.813269 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 05:43:02.813282 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.266667 | |
| I0408 05:43:02.813299 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.0193 (* 0.3 = 0.905791 loss) | |
| I0408 05:43:02.813315 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.82371 (* 0.3 = 0.247113 loss) | |
| I0408 05:43:02.813328 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.133333 | |
| I0408 05:43:02.813340 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 05:43:02.813354 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.377778 | |
| I0408 05:43:02.813369 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.89512 (* 0.3 = 0.868535 loss) | |
| I0408 05:43:02.813383 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.78978 (* 0.3 = 0.236934 loss) | |
| I0408 05:43:02.813395 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.333333 | |
| I0408 05:43:02.813408 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 05:43:02.813421 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.622222 | |
| I0408 05:43:02.813436 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.54797 (* 1 = 2.54797 loss) | |
| I0408 05:43:02.813449 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.704613 (* 1 = 0.704613 loss) | |
| I0408 05:43:02.813462 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:43:02.813477 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00452278 | |
| I0408 05:43:02.813491 3443 sgd_solver.cpp:106] Iteration 31000, lr = 0.00955714 | |
| I0408 05:48:36.197264 3443 solver.cpp:229] Iteration 31500, loss = 5.10699 | |
| I0408 05:48:36.197386 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1 | |
| I0408 05:48:36.197405 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 05:48:36.197419 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.3 | |
| I0408 05:48:36.197437 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.78707 (* 0.3 = 0.83612 loss) | |
| I0408 05:48:36.197453 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.821366 (* 0.3 = 0.24641 loss) | |
| I0408 05:48:36.197465 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.18 | |
| I0408 05:48:36.197479 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.767045 | |
| I0408 05:48:36.197490 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.4 | |
| I0408 05:48:36.197505 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.52776 (* 0.3 = 0.758329 loss) | |
| I0408 05:48:36.197520 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.742024 (* 0.3 = 0.222607 loss) | |
| I0408 05:48:36.197532 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.32 | |
| I0408 05:48:36.197546 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.801136 | |
| I0408 05:48:36.197561 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.7 | |
| I0408 05:48:36.197576 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.08504 (* 1 = 2.08504 loss) | |
| I0408 05:48:36.197590 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.63094 (* 1 = 0.63094 loss) | |
| I0408 05:48:36.197603 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:48:36.197615 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00303106 | |
| I0408 05:48:36.197629 3443 sgd_solver.cpp:106] Iteration 31500, lr = 0.00955 | |
| I0408 05:54:09.581734 3443 solver.cpp:229] Iteration 32000, loss = 5.04984 | |
| I0408 05:54:09.581858 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0612245 | |
| I0408 05:54:09.581879 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 05:54:09.581893 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.244898 | |
| I0408 05:54:09.581923 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.16992 (* 0.3 = 0.950975 loss) | |
| I0408 05:54:09.581939 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.941856 (* 0.3 = 0.282557 loss) | |
| I0408 05:54:09.581953 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.163265 | |
| I0408 05:54:09.581965 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 05:54:09.581985 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.326531 | |
| I0408 05:54:09.582000 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.96188 (* 0.3 = 0.888563 loss) | |
| I0408 05:54:09.582015 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.871674 (* 0.3 = 0.261502 loss) | |
| I0408 05:54:09.582027 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.22449 | |
| I0408 05:54:09.582039 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.778409 | |
| I0408 05:54:09.582052 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.510204 | |
| I0408 05:54:09.582067 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.5644 (* 1 = 2.5644 loss) | |
| I0408 05:54:09.582082 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.748741 (* 1 = 0.748741 loss) | |
| I0408 05:54:09.582093 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:54:09.582106 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00474768 | |
| I0408 05:54:09.582123 3443 sgd_solver.cpp:106] Iteration 32000, lr = 0.00954286 | |
| I0408 05:59:42.964702 3443 solver.cpp:229] Iteration 32500, loss = 5.04794 | |
| I0408 05:59:42.964870 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170213 | |
| I0408 05:59:42.964891 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 05:59:42.964905 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.340426 | |
| I0408 05:59:42.964923 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.97821 (* 0.3 = 0.893463 loss) | |
| I0408 05:59:42.964939 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.888276 (* 0.3 = 0.266483 loss) | |
| I0408 05:59:42.964962 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.234043 | |
| I0408 05:59:42.964974 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 05:59:42.964987 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.446809 | |
| I0408 05:59:42.965000 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.45597 (* 0.3 = 0.736791 loss) | |
| I0408 05:59:42.965016 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.801562 (* 0.3 = 0.240469 loss) | |
| I0408 05:59:42.965037 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.404255 | |
| I0408 05:59:42.965049 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 05:59:42.965062 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.617021 | |
| I0408 05:59:42.965076 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.95904 (* 1 = 1.95904 loss) | |
| I0408 05:59:42.965098 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.567995 (* 1 = 0.567995 loss) | |
| I0408 05:59:42.965111 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 05:59:42.965123 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0299854 | |
| I0408 05:59:42.965138 3443 sgd_solver.cpp:106] Iteration 32500, lr = 0.00953571 | |
| I0408 06:05:16.352061 3443 solver.cpp:229] Iteration 33000, loss = 4.97942 | |
| I0408 06:05:16.352185 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.265306 | |
| I0408 06:05:16.352206 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 06:05:16.352221 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.489796 | |
| I0408 06:05:16.352237 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.67611 (* 0.3 = 0.802834 loss) | |
| I0408 06:05:16.352253 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.826175 (* 0.3 = 0.247852 loss) | |
| I0408 06:05:16.352265 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.265306 | |
| I0408 06:05:16.352279 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 06:05:16.352293 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.469388 | |
| I0408 06:05:16.352306 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.50913 (* 0.3 = 0.752738 loss) | |
| I0408 06:05:16.352320 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.752596 (* 0.3 = 0.225779 loss) | |
| I0408 06:05:16.352334 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.326531 | |
| I0408 06:05:16.352345 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.795455 | |
| I0408 06:05:16.352357 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.693878 | |
| I0408 06:05:16.352373 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.93327 (* 1 = 1.93327 loss) | |
| I0408 06:05:16.352387 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.587124 (* 1 = 0.587124 loss) | |
| I0408 06:05:16.352401 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:05:16.352412 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0250051 | |
| I0408 06:05:16.352427 3443 sgd_solver.cpp:106] Iteration 33000, lr = 0.00952857 | |
| I0408 06:10:49.742142 3443 solver.cpp:229] Iteration 33500, loss = 4.93185 | |
| I0408 06:10:49.742288 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.139535 | |
| I0408 06:10:49.742310 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 06:10:49.742323 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.325581 | |
| I0408 06:10:49.742341 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.62702 (* 0.3 = 0.788106 loss) | |
| I0408 06:10:49.742355 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.757426 (* 0.3 = 0.227228 loss) | |
| I0408 06:10:49.742368 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.27907 | |
| I0408 06:10:49.742382 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 06:10:49.742393 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.581395 | |
| I0408 06:10:49.742408 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.32824 (* 0.3 = 0.698473 loss) | |
| I0408 06:10:49.742422 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.670243 (* 0.3 = 0.201073 loss) | |
| I0408 06:10:49.742435 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.372093 | |
| I0408 06:10:49.742447 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.818182 | |
| I0408 06:10:49.742460 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.697674 | |
| I0408 06:10:49.742475 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.86117 (* 1 = 1.86117 loss) | |
| I0408 06:10:49.742488 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.543658 (* 1 = 0.543658 loss) | |
| I0408 06:10:49.742501 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:10:49.742513 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00785647 | |
| I0408 06:10:49.742527 3443 sgd_solver.cpp:106] Iteration 33500, lr = 0.00952143 | |
| I0408 06:16:23.122870 3443 solver.cpp:229] Iteration 34000, loss = 4.84969 | |
| I0408 06:16:23.122992 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.230769 | |
| I0408 06:16:23.123013 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 06:16:23.123025 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.403846 | |
| I0408 06:16:23.123044 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.88372 (* 0.3 = 0.865115 loss) | |
| I0408 06:16:23.123059 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.877545 (* 0.3 = 0.263263 loss) | |
| I0408 06:16:23.123072 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.288462 | |
| I0408 06:16:23.123085 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 06:16:23.123097 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.5 | |
| I0408 06:16:23.123111 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.48519 (* 0.3 = 0.745559 loss) | |
| I0408 06:16:23.123126 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.750957 (* 0.3 = 0.225287 loss) | |
| I0408 06:16:23.123139 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.288462 | |
| I0408 06:16:23.123152 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.784091 | |
| I0408 06:16:23.123163 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.576923 | |
| I0408 06:16:23.123178 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.23586 (* 1 = 2.23586 loss) | |
| I0408 06:16:23.123193 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.679149 (* 1 = 0.679149 loss) | |
| I0408 06:16:23.123204 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:16:23.123216 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00947154 | |
| I0408 06:16:23.123231 3443 sgd_solver.cpp:106] Iteration 34000, lr = 0.00951429 | |
| I0408 06:21:56.498411 3443 solver.cpp:229] Iteration 34500, loss = 4.85039 | |
| I0408 06:21:56.498551 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.131579 | |
| I0408 06:21:56.498572 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 06:21:56.498586 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.289474 | |
| I0408 06:21:56.498605 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.85295 (* 0.3 = 0.855886 loss) | |
| I0408 06:21:56.498620 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.733991 (* 0.3 = 0.220197 loss) | |
| I0408 06:21:56.498633 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.210526 | |
| I0408 06:21:56.498646 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 06:21:56.498661 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.5 | |
| I0408 06:21:56.498674 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.47461 (* 0.3 = 0.742383 loss) | |
| I0408 06:21:56.498689 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.643811 (* 0.3 = 0.193143 loss) | |
| I0408 06:21:56.498703 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.315789 | |
| I0408 06:21:56.498715 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.829545 | |
| I0408 06:21:56.498728 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.684211 | |
| I0408 06:21:56.498742 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.07816 (* 1 = 2.07816 loss) | |
| I0408 06:21:56.498759 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.547755 (* 1 = 0.547755 loss) | |
| I0408 06:21:56.498770 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:21:56.498782 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00605544 | |
| I0408 06:21:56.498797 3443 sgd_solver.cpp:106] Iteration 34500, lr = 0.00950714 | |
| I0408 06:27:29.485337 3443 solver.cpp:338] Iteration 35000, Testing net (#0) | |
| I0408 06:28:10.367095 3443 solver.cpp:393] Test loss: 4.0954 | |
| I0408 06:28:10.367198 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.215172 | |
| I0408 06:28:10.367218 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.805272 | |
| I0408 06:28:10.367233 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.471191 | |
| I0408 06:28:10.367249 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.61869 (* 0.3 = 0.785608 loss) | |
| I0408 06:28:10.367264 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.669698 (* 0.3 = 0.200909 loss) | |
| I0408 06:28:10.367276 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.299377 | |
| I0408 06:28:10.367288 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.821319 | |
| I0408 06:28:10.367300 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.613951 | |
| I0408 06:28:10.367314 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 2.24778 (* 0.3 = 0.674333 loss) | |
| I0408 06:28:10.367346 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.590392 (* 0.3 = 0.177118 loss) | |
| I0408 06:28:10.367360 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.451047 | |
| I0408 06:28:10.367372 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.854503 | |
| I0408 06:28:10.367384 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.747107 | |
| I0408 06:28:10.367398 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.77778 (* 1 = 1.77778 loss) | |
| I0408 06:28:10.367413 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.479648 (* 1 = 0.479648 loss) | |
| I0408 06:28:10.367424 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.017 | |
| I0408 06:28:10.367436 3443 solver.cpp:406] Test net output #16: total_confidence = 0.0301474 | |
| I0408 06:28:10.741078 3443 solver.cpp:229] Iteration 35000, loss = 4.83928 | |
| I0408 06:28:10.741122 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.148936 | |
| I0408 06:28:10.741139 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 06:28:10.741153 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.276596 | |
| I0408 06:28:10.741168 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.1082 (* 0.3 = 0.932461 loss) | |
| I0408 06:28:10.741184 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.890236 (* 0.3 = 0.267071 loss) | |
| I0408 06:28:10.741196 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.212766 | |
| I0408 06:28:10.741209 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 06:28:10.741221 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.468085 | |
| I0408 06:28:10.741236 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.76517 (* 0.3 = 0.82955 loss) | |
| I0408 06:28:10.741251 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.794256 (* 0.3 = 0.238277 loss) | |
| I0408 06:28:10.741263 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.276596 | |
| I0408 06:28:10.741276 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.801136 | |
| I0408 06:28:10.741287 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.595745 | |
| I0408 06:28:10.741302 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.28737 (* 1 = 2.28737 loss) | |
| I0408 06:28:10.741315 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.635768 (* 1 = 0.635768 loss) | |
| I0408 06:28:10.741336 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:28:10.741359 3443 solver.cpp:245] Train net output #16: total_confidence = 0.035943 | |
| I0408 06:28:10.741376 3443 sgd_solver.cpp:106] Iteration 35000, lr = 0.0095 | |
| I0408 06:33:43.934198 3443 solver.cpp:229] Iteration 35500, loss = 4.75338 | |
| I0408 06:33:43.934358 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.0652174 | |
| I0408 06:33:43.934381 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 06:33:43.934393 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.369565 | |
| I0408 06:33:43.934411 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.21137 (* 0.3 = 0.96341 loss) | |
| I0408 06:33:43.934427 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.944004 (* 0.3 = 0.283201 loss) | |
| I0408 06:33:43.934439 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.173913 | |
| I0408 06:33:43.934451 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 06:33:43.934463 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.434783 | |
| I0408 06:33:43.934478 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.12269 (* 0.3 = 0.936808 loss) | |
| I0408 06:33:43.934492 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.912652 (* 0.3 = 0.273796 loss) | |
| I0408 06:33:43.934504 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.369565 | |
| I0408 06:33:43.934516 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 06:33:43.934530 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.543478 | |
| I0408 06:33:43.934545 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.24622 (* 1 = 2.24622 loss) | |
| I0408 06:33:43.934558 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.654418 (* 1 = 0.654418 loss) | |
| I0408 06:33:43.934571 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 06:33:43.934583 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0109744 | |
| I0408 06:33:43.934597 3443 sgd_solver.cpp:106] Iteration 35500, lr = 0.00949286 | |
| I0408 06:39:17.318689 3443 solver.cpp:229] Iteration 36000, loss = 4.81359 | |
| I0408 06:39:17.318843 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.181818 | |
| I0408 06:39:17.318864 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 06:39:17.318878 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.454545 | |
| I0408 06:39:17.318895 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.80751 (* 0.3 = 0.842254 loss) | |
| I0408 06:39:17.318910 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.804576 (* 0.3 = 0.241373 loss) | |
| I0408 06:39:17.318927 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.272727 | |
| I0408 06:39:17.318939 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 06:39:17.318953 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.590909 | |
| I0408 06:39:17.318967 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.44805 (* 0.3 = 0.734414 loss) | |
| I0408 06:39:17.318981 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.746742 (* 0.3 = 0.224023 loss) | |
| I0408 06:39:17.318994 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.386364 | |
| I0408 06:39:17.319007 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.840909 | |
| I0408 06:39:17.319020 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.75 | |
| I0408 06:39:17.319034 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.73737 (* 1 = 1.73737 loss) | |
| I0408 06:39:17.319048 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.468786 (* 1 = 0.468786 loss) | |
| I0408 06:39:17.319061 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:39:17.319073 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0451285 | |
| I0408 06:39:17.319087 3443 sgd_solver.cpp:106] Iteration 36000, lr = 0.00948571 | |
| I0408 06:44:50.705765 3443 solver.cpp:229] Iteration 36500, loss = 4.65573 | |
| I0408 06:44:50.705905 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1875 | |
| I0408 06:44:50.705929 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 06:44:50.705942 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.416667 | |
| I0408 06:44:50.705960 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.83622 (* 0.3 = 0.850867 loss) | |
| I0408 06:44:50.705976 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.828629 (* 0.3 = 0.248589 loss) | |
| I0408 06:44:50.705987 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.208333 | |
| I0408 06:44:50.706001 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 06:44:50.706013 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.4375 | |
| I0408 06:44:50.706027 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.69789 (* 0.3 = 0.809367 loss) | |
| I0408 06:44:50.706043 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.797178 (* 0.3 = 0.239154 loss) | |
| I0408 06:44:50.706056 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.375 | |
| I0408 06:44:50.706068 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.818182 | |
| I0408 06:44:50.706080 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.729167 | |
| I0408 06:44:50.706094 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.95794 (* 1 = 1.95794 loss) | |
| I0408 06:44:50.706110 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.574129 (* 1 = 0.574129 loss) | |
| I0408 06:44:50.706121 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:44:50.706133 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0182024 | |
| I0408 06:44:50.706148 3443 sgd_solver.cpp:106] Iteration 36500, lr = 0.00947857 | |
| I0408 06:50:24.272944 3443 solver.cpp:229] Iteration 37000, loss = 4.70395 | |
| I0408 06:50:24.273102 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.232558 | |
| I0408 06:50:24.273134 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 06:50:24.273159 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.511628 | |
| I0408 06:50:24.273185 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.55963 (* 0.3 = 0.767889 loss) | |
| I0408 06:50:24.273216 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.667124 (* 0.3 = 0.200137 loss) | |
| I0408 06:50:24.273233 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.372093 | |
| I0408 06:50:24.273247 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 06:50:24.273259 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.604651 | |
| I0408 06:50:24.273275 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.22153 (* 0.3 = 0.66646 loss) | |
| I0408 06:50:24.273290 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.603912 (* 0.3 = 0.181174 loss) | |
| I0408 06:50:24.273303 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.395349 | |
| I0408 06:50:24.273315 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.846591 | |
| I0408 06:50:24.273327 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.837209 | |
| I0408 06:50:24.273344 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.61394 (* 1 = 1.61394 loss) | |
| I0408 06:50:24.273357 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.425367 (* 1 = 0.425367 loss) | |
| I0408 06:50:24.273370 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:50:24.273382 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0102365 | |
| I0408 06:50:24.273398 3443 sgd_solver.cpp:106] Iteration 37000, lr = 0.00947143 | |
| I0408 06:55:57.808449 3443 solver.cpp:229] Iteration 37500, loss = 4.66713 | |
| I0408 06:55:57.808560 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.181818 | |
| I0408 06:55:57.808579 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 06:55:57.808593 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.431818 | |
| I0408 06:55:57.808611 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.81697 (* 0.3 = 0.84509 loss) | |
| I0408 06:55:57.808627 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.785185 (* 0.3 = 0.235555 loss) | |
| I0408 06:55:57.808640 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.295455 | |
| I0408 06:55:57.808652 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 06:55:57.808665 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.590909 | |
| I0408 06:55:57.808679 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.42629 (* 0.3 = 0.727886 loss) | |
| I0408 06:55:57.808693 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.669031 (* 0.3 = 0.200709 loss) | |
| I0408 06:55:57.808706 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.568182 | |
| I0408 06:55:57.808719 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 06:55:57.808732 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.818182 | |
| I0408 06:55:57.808746 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.56279 (* 1 = 1.56279 loss) | |
| I0408 06:55:57.808760 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.453297 (* 1 = 0.453297 loss) | |
| I0408 06:55:57.808773 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 06:55:57.808784 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0321454 | |
| I0408 06:55:57.808799 3443 sgd_solver.cpp:106] Iteration 37500, lr = 0.00946429 | |
| I0408 07:01:31.207726 3443 solver.cpp:229] Iteration 38000, loss = 4.63113 | |
| I0408 07:01:31.207869 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.28 | |
| I0408 07:01:31.207900 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 07:01:31.207926 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.58 | |
| I0408 07:01:31.207957 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.39881 (* 0.3 = 0.719643 loss) | |
| I0408 07:01:31.207984 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.728012 (* 0.3 = 0.218404 loss) | |
| I0408 07:01:31.208006 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.26 | |
| I0408 07:01:31.208027 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 07:01:31.208048 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.64 | |
| I0408 07:01:31.208076 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.15989 (* 0.3 = 0.647967 loss) | |
| I0408 07:01:31.208103 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.641307 (* 0.3 = 0.192392 loss) | |
| I0408 07:01:31.208124 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.4 | |
| I0408 07:01:31.208145 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 07:01:31.208165 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.72 | |
| I0408 07:01:31.208191 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.75136 (* 1 = 1.75136 loss) | |
| I0408 07:01:31.208219 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.516362 (* 1 = 0.516362 loss) | |
| I0408 07:01:31.208242 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:01:31.208263 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0133313 | |
| I0408 07:01:31.208287 3443 sgd_solver.cpp:106] Iteration 38000, lr = 0.00945714 | |
| I0408 07:07:04.594871 3443 solver.cpp:229] Iteration 38500, loss = 4.5094 | |
| I0408 07:07:04.595007 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.26 | |
| I0408 07:07:04.595038 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 07:07:04.595060 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.4 | |
| I0408 07:07:04.595090 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.6328 (* 0.3 = 0.789839 loss) | |
| I0408 07:07:04.595119 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.791143 (* 0.3 = 0.237343 loss) | |
| I0408 07:07:04.595140 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.32 | |
| I0408 07:07:04.595163 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 07:07:04.595185 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.56 | |
| I0408 07:07:04.595211 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.31136 (* 0.3 = 0.693408 loss) | |
| I0408 07:07:04.595239 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.712426 (* 0.3 = 0.213728 loss) | |
| I0408 07:07:04.595262 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.44 | |
| I0408 07:07:04.595283 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.840909 | |
| I0408 07:07:04.595304 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.72 | |
| I0408 07:07:04.595351 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.63677 (* 1 = 1.63677 loss) | |
| I0408 07:07:04.595381 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.503981 (* 1 = 0.503981 loss) | |
| I0408 07:07:04.595405 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:07:04.595428 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0520383 | |
| I0408 07:07:04.595454 3443 sgd_solver.cpp:106] Iteration 38500, lr = 0.00945 | |
| I0408 07:12:37.969761 3443 solver.cpp:229] Iteration 39000, loss = 4.59841 | |
| I0408 07:12:37.969902 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.181818 | |
| I0408 07:12:37.969925 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 07:12:37.969938 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.386364 | |
| I0408 07:12:37.969955 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.92327 (* 0.3 = 0.876981 loss) | |
| I0408 07:12:37.969970 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.787195 (* 0.3 = 0.236158 loss) | |
| I0408 07:12:37.969983 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.295455 | |
| I0408 07:12:37.969995 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0408 07:12:37.970008 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.522727 | |
| I0408 07:12:37.970022 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.62766 (* 0.3 = 0.788298 loss) | |
| I0408 07:12:37.970037 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.718012 (* 0.3 = 0.215404 loss) | |
| I0408 07:12:37.970051 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.454545 | |
| I0408 07:12:37.970062 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.846591 | |
| I0408 07:12:37.970074 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.636364 | |
| I0408 07:12:37.970089 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.16104 (* 1 = 2.16104 loss) | |
| I0408 07:12:37.970103 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.584129 (* 1 = 0.584129 loss) | |
| I0408 07:12:37.970115 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:12:37.970127 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0392059 | |
| I0408 07:12:37.970142 3443 sgd_solver.cpp:106] Iteration 39000, lr = 0.00944286 | |
| I0408 07:18:11.350213 3443 solver.cpp:229] Iteration 39500, loss = 4.58564 | |
| I0408 07:18:11.350385 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.261905 | |
| I0408 07:18:11.350407 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 07:18:11.350421 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.380952 | |
| I0408 07:18:11.350440 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.63829 (* 0.3 = 0.791486 loss) | |
| I0408 07:18:11.350455 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.720649 (* 0.3 = 0.216195 loss) | |
| I0408 07:18:11.350467 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.190476 | |
| I0408 07:18:11.350481 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 07:18:11.350493 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.47619 | |
| I0408 07:18:11.350509 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.46013 (* 0.3 = 0.738039 loss) | |
| I0408 07:18:11.350524 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.694471 (* 0.3 = 0.208341 loss) | |
| I0408 07:18:11.350536 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.52381 | |
| I0408 07:18:11.350549 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 07:18:11.350563 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.833333 | |
| I0408 07:18:11.350576 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.52654 (* 1 = 1.52654 loss) | |
| I0408 07:18:11.350591 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.420523 (* 1 = 0.420523 loss) | |
| I0408 07:18:11.350605 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:18:11.350616 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0155166 | |
| I0408 07:18:11.350631 3443 sgd_solver.cpp:106] Iteration 39500, lr = 0.00943571 | |
| I0408 07:23:44.363075 3443 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_40000.caffemodel | |
| I0408 07:23:44.896877 3443 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_40000.solverstate | |
| I0408 07:23:45.160977 3443 solver.cpp:338] Iteration 40000, Testing net (#0) | |
| I0408 07:24:25.984428 3443 solver.cpp:393] Test loss: 4.28317 | |
| I0408 07:24:25.984539 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.201693 | |
| I0408 07:24:25.984558 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.792636 | |
| I0408 07:24:25.984572 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.435667 | |
| I0408 07:24:25.984589 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.70979 (* 0.3 = 0.812938 loss) | |
| I0408 07:24:25.984603 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.737151 (* 0.3 = 0.221145 loss) | |
| I0408 07:24:25.984616 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.31752 | |
| I0408 07:24:25.984628 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.824 | |
| I0408 07:24:25.984642 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.618525 | |
| I0408 07:24:25.984655 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 2.34303 (* 0.3 = 0.702908 loss) | |
| I0408 07:24:25.984669 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.617507 (* 0.3 = 0.185252 loss) | |
| I0408 07:24:25.984681 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.466562 | |
| I0408 07:24:25.984694 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.860321 | |
| I0408 07:24:25.984705 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.736778 | |
| I0408 07:24:25.984719 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.87025 (* 1 = 1.87025 loss) | |
| I0408 07:24:25.984733 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.490685 (* 1 = 0.490685 loss) | |
| I0408 07:24:25.984745 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.025 | |
| I0408 07:24:25.984757 3443 solver.cpp:406] Test net output #16: total_confidence = 0.0479606 | |
| I0408 07:24:26.357055 3443 solver.cpp:229] Iteration 40000, loss = 4.51147 | |
| I0408 07:24:26.357100 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.214286 | |
| I0408 07:24:26.357120 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 07:24:26.357132 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.47619 | |
| I0408 07:24:26.357148 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.51309 (* 0.3 = 0.753927 loss) | |
| I0408 07:24:26.357163 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.682643 (* 0.3 = 0.204793 loss) | |
| I0408 07:24:26.357177 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.452381 | |
| I0408 07:24:26.357189 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0408 07:24:26.357202 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 07:24:26.357215 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.91446 (* 0.3 = 0.574338 loss) | |
| I0408 07:24:26.357230 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.526834 (* 0.3 = 0.15805 loss) | |
| I0408 07:24:26.357244 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.428571 | |
| I0408 07:24:26.357255 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.852273 | |
| I0408 07:24:26.357267 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.761905 | |
| I0408 07:24:26.357282 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.51038 (* 1 = 1.51038 loss) | |
| I0408 07:24:26.357296 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.430583 (* 1 = 0.430583 loss) | |
| I0408 07:24:26.357308 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 07:24:26.357321 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0117866 | |
| I0408 07:24:26.357336 3443 sgd_solver.cpp:106] Iteration 40000, lr = 0.00942857 | |
| I0408 07:29:59.787551 3443 solver.cpp:229] Iteration 40500, loss = 4.53543 | |
| I0408 07:29:59.787688 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.159091 | |
| I0408 07:29:59.787708 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 07:29:59.787721 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.454545 | |
| I0408 07:29:59.787737 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.74372 (* 0.3 = 0.823117 loss) | |
| I0408 07:29:59.787753 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.996515 (* 0.3 = 0.298955 loss) | |
| I0408 07:29:59.787766 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.25 | |
| I0408 07:29:59.787778 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 07:29:59.787791 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.522727 | |
| I0408 07:29:59.787806 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.46781 (* 0.3 = 0.740344 loss) | |
| I0408 07:29:59.787820 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.857632 (* 0.3 = 0.25729 loss) | |
| I0408 07:29:59.787832 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.409091 | |
| I0408 07:29:59.787845 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.801136 | |
| I0408 07:29:59.787858 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.704545 | |
| I0408 07:29:59.787871 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.0294 (* 1 = 2.0294 loss) | |
| I0408 07:29:59.787886 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.764169 (* 1 = 0.764169 loss) | |
| I0408 07:29:59.787899 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 07:29:59.787910 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0122209 | |
| I0408 07:29:59.787927 3443 sgd_solver.cpp:106] Iteration 40500, lr = 0.00942143 | |
| I0408 07:35:33.181404 3443 solver.cpp:229] Iteration 41000, loss = 4.38253 | |
| I0408 07:35:33.181511 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.179487 | |
| I0408 07:35:33.181530 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 07:35:33.181545 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.487179 | |
| I0408 07:35:33.181561 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.73064 (* 0.3 = 0.819191 loss) | |
| I0408 07:35:33.181577 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.718487 (* 0.3 = 0.215546 loss) | |
| I0408 07:35:33.181591 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.307692 | |
| I0408 07:35:33.181602 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 07:35:33.181615 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.564103 | |
| I0408 07:35:33.181629 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.3394 (* 0.3 = 0.70182 loss) | |
| I0408 07:35:33.181644 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.636151 (* 0.3 = 0.190845 loss) | |
| I0408 07:35:33.181656 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.512821 | |
| I0408 07:35:33.181669 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 07:35:33.181681 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.794872 | |
| I0408 07:35:33.181695 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.45096 (* 1 = 1.45096 loss) | |
| I0408 07:35:33.181710 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.37363 (* 1 = 0.37363 loss) | |
| I0408 07:35:33.181722 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:35:33.181733 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0165384 | |
| I0408 07:35:33.181748 3443 sgd_solver.cpp:106] Iteration 41000, lr = 0.00941429 | |
| I0408 07:41:06.555713 3443 solver.cpp:229] Iteration 41500, loss = 4.33516 | |
| I0408 07:41:06.555891 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1875 | |
| I0408 07:41:06.555912 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 07:41:06.555929 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.333333 | |
| I0408 07:41:06.555948 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.79003 (* 0.3 = 0.837008 loss) | |
| I0408 07:41:06.555963 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.80129 (* 0.3 = 0.240387 loss) | |
| I0408 07:41:06.555976 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.1875 | |
| I0408 07:41:06.555989 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.755682 | |
| I0408 07:41:06.556002 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.541667 | |
| I0408 07:41:06.556016 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.63525 (* 0.3 = 0.790576 loss) | |
| I0408 07:41:06.556031 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.803931 (* 0.3 = 0.241179 loss) | |
| I0408 07:41:06.556043 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.458333 | |
| I0408 07:41:06.556056 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 07:41:06.556068 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.75 | |
| I0408 07:41:06.556083 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.73477 (* 1 = 1.73477 loss) | |
| I0408 07:41:06.556098 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.533991 (* 1 = 0.533991 loss) | |
| I0408 07:41:06.556110 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:41:06.556123 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0100958 | |
| I0408 07:41:06.556138 3443 sgd_solver.cpp:106] Iteration 41500, lr = 0.00940714 | |
| I0408 07:46:39.958488 3443 solver.cpp:229] Iteration 42000, loss = 4.41297 | |
| I0408 07:46:39.958622 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.108696 | |
| I0408 07:46:39.958643 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 07:46:39.958657 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.173913 | |
| I0408 07:46:39.958673 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.99757 (* 0.3 = 0.899272 loss) | |
| I0408 07:46:39.958689 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.927949 (* 0.3 = 0.278385 loss) | |
| I0408 07:46:39.958703 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.195652 | |
| I0408 07:46:39.958715 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 07:46:39.958727 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.434783 | |
| I0408 07:46:39.958742 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.74967 (* 0.3 = 0.8249 loss) | |
| I0408 07:46:39.958756 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.84251 (* 0.3 = 0.252753 loss) | |
| I0408 07:46:39.958770 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.434783 | |
| I0408 07:46:39.958782 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.835227 | |
| I0408 07:46:39.958796 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.630435 | |
| I0408 07:46:39.958811 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.94645 (* 1 = 1.94645 loss) | |
| I0408 07:46:39.958824 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.598705 (* 1 = 0.598705 loss) | |
| I0408 07:46:39.958837 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:46:39.958848 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00729562 | |
| I0408 07:46:39.958863 3443 sgd_solver.cpp:106] Iteration 42000, lr = 0.0094 | |
| I0408 07:52:14.001816 3443 solver.cpp:229] Iteration 42500, loss = 4.32487 | |
| I0408 07:52:14.001966 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.204545 | |
| I0408 07:52:14.001987 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 07:52:14.002002 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.5 | |
| I0408 07:52:14.002019 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.52528 (* 0.3 = 0.757583 loss) | |
| I0408 07:52:14.002034 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.732987 (* 0.3 = 0.219896 loss) | |
| I0408 07:52:14.002048 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.386364 | |
| I0408 07:52:14.002060 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 07:52:14.002073 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.636364 | |
| I0408 07:52:14.002087 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.14141 (* 0.3 = 0.642424 loss) | |
| I0408 07:52:14.002102 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.632104 (* 0.3 = 0.189631 loss) | |
| I0408 07:52:14.002115 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.477273 | |
| I0408 07:52:14.002128 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 07:52:14.002140 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.795455 | |
| I0408 07:52:14.002154 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.43281 (* 1 = 1.43281 loss) | |
| I0408 07:52:14.002168 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.397218 (* 1 = 0.397218 loss) | |
| I0408 07:52:14.002182 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:52:14.002193 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0298337 | |
| I0408 07:52:14.002209 3443 sgd_solver.cpp:106] Iteration 42500, lr = 0.00939286 | |
| I0408 07:57:47.741140 3443 solver.cpp:229] Iteration 43000, loss = 4.30786 | |
| I0408 07:57:47.741315 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.270833 | |
| I0408 07:57:47.741338 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 07:57:47.741351 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.520833 | |
| I0408 07:57:47.741370 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.52953 (* 0.3 = 0.75886 loss) | |
| I0408 07:57:47.741390 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.731093 (* 0.3 = 0.219328 loss) | |
| I0408 07:57:47.741418 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.458333 | |
| I0408 07:57:47.741446 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0408 07:57:47.741475 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 07:57:47.741509 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.25857 (* 0.3 = 0.677571 loss) | |
| I0408 07:57:47.741544 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.643874 (* 0.3 = 0.193162 loss) | |
| I0408 07:57:47.741574 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.583333 | |
| I0408 07:57:47.741601 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0408 07:57:47.741629 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.895833 | |
| I0408 07:57:47.741662 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.31631 (* 1 = 1.31631 loss) | |
| I0408 07:57:47.741691 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.373632 (* 1 = 0.373632 loss) | |
| I0408 07:57:47.741717 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 07:57:47.741739 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0408934 | |
| I0408 07:57:47.741761 3443 sgd_solver.cpp:106] Iteration 43000, lr = 0.00938571 | |
| I0408 08:03:21.110702 3443 solver.cpp:229] Iteration 43500, loss = 4.29116 | |
| I0408 08:03:21.110875 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.196078 | |
| I0408 08:03:21.110898 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 08:03:21.110911 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.45098 | |
| I0408 08:03:21.110931 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.9299 (* 0.3 = 0.878971 loss) | |
| I0408 08:03:21.110946 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.927684 (* 0.3 = 0.278305 loss) | |
| I0408 08:03:21.110960 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.333333 | |
| I0408 08:03:21.110977 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 08:03:21.110991 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.568627 | |
| I0408 08:03:21.111006 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.44474 (* 0.3 = 0.733423 loss) | |
| I0408 08:03:21.111021 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.76311 (* 0.3 = 0.228933 loss) | |
| I0408 08:03:21.111033 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.470588 | |
| I0408 08:03:21.111047 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.840909 | |
| I0408 08:03:21.111058 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.745098 | |
| I0408 08:03:21.111073 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.74062 (* 1 = 1.74062 loss) | |
| I0408 08:03:21.111088 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.526914 (* 1 = 0.526914 loss) | |
| I0408 08:03:21.111101 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:03:21.111114 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0392329 | |
| I0408 08:03:21.111129 3443 sgd_solver.cpp:106] Iteration 43500, lr = 0.00937857 | |
| I0408 08:08:54.502329 3443 solver.cpp:229] Iteration 44000, loss = 4.26276 | |
| I0408 08:08:54.502465 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.1875 | |
| I0408 08:08:54.502485 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 08:08:54.502501 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.270833 | |
| I0408 08:08:54.502517 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.2683 (* 0.3 = 0.98049 loss) | |
| I0408 08:08:54.502533 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.978263 (* 0.3 = 0.293479 loss) | |
| I0408 08:08:54.502547 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.25 | |
| I0408 08:08:54.502560 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 08:08:54.502573 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.4375 | |
| I0408 08:08:54.502588 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.99348 (* 0.3 = 0.898043 loss) | |
| I0408 08:08:54.502602 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.861443 (* 0.3 = 0.258433 loss) | |
| I0408 08:08:54.502615 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.333333 | |
| I0408 08:08:54.502627 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.818182 | |
| I0408 08:08:54.502640 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.666667 | |
| I0408 08:08:54.502655 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.07118 (* 1 = 2.07118 loss) | |
| I0408 08:08:54.502670 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.588973 (* 1 = 0.588973 loss) | |
| I0408 08:08:54.502682 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:08:54.502694 3443 solver.cpp:245] Train net output #16: total_confidence = 0.00549494 | |
| I0408 08:08:54.502710 3443 sgd_solver.cpp:106] Iteration 44000, lr = 0.00937143 | |
| I0408 08:14:27.913336 3443 solver.cpp:229] Iteration 44500, loss = 4.18289 | |
| I0408 08:14:27.913528 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.229167 | |
| I0408 08:14:27.913550 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 08:14:27.913573 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.4375 | |
| I0408 08:14:27.913590 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.91427 (* 0.3 = 0.874283 loss) | |
| I0408 08:14:27.913605 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.869366 (* 0.3 = 0.26081 loss) | |
| I0408 08:14:27.913619 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.229167 | |
| I0408 08:14:27.913632 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 08:14:27.913645 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.5625 | |
| I0408 08:14:27.913660 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.59131 (* 0.3 = 0.777394 loss) | |
| I0408 08:14:27.913674 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.77895 (* 0.3 = 0.233685 loss) | |
| I0408 08:14:27.913687 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.5 | |
| I0408 08:14:27.913700 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 08:14:27.913712 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.770833 | |
| I0408 08:14:27.913728 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.73475 (* 1 = 1.73475 loss) | |
| I0408 08:14:27.913743 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.509851 (* 1 = 0.509851 loss) | |
| I0408 08:14:27.913755 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:14:27.913769 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0273994 | |
| I0408 08:14:27.913784 3443 sgd_solver.cpp:106] Iteration 44500, lr = 0.00936429 | |
| I0408 08:20:00.877070 3443 solver.cpp:338] Iteration 45000, Testing net (#0) | |
| I0408 08:20:42.370219 3443 solver.cpp:393] Test loss: 3.70414 | |
| I0408 08:20:42.370339 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.262795 | |
| I0408 08:20:42.370358 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.812455 | |
| I0408 08:20:42.370373 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.561312 | |
| I0408 08:20:42.370390 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.33902 (* 0.3 = 0.701707 loss) | |
| I0408 08:20:42.370406 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.62291 (* 0.3 = 0.186873 loss) | |
| I0408 08:20:42.370419 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.406812 | |
| I0408 08:20:42.370431 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.844274 | |
| I0408 08:20:42.370443 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.712847 | |
| I0408 08:20:42.370458 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.96986 (* 0.3 = 0.590959 loss) | |
| I0408 08:20:42.370472 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.535288 (* 0.3 = 0.160586 loss) | |
| I0408 08:20:42.370486 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.515773 | |
| I0408 08:20:42.370497 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.879186 | |
| I0408 08:20:42.370509 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.811592 | |
| I0408 08:20:42.370523 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.64699 (* 1 = 1.64699 loss) | |
| I0408 08:20:42.370537 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.417033 (* 1 = 0.417033 loss) | |
| I0408 08:20:42.370550 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.057 | |
| I0408 08:20:42.370563 3443 solver.cpp:406] Test net output #16: total_confidence = 0.0777983 | |
| I0408 08:20:42.749699 3443 solver.cpp:229] Iteration 45000, loss = 4.22782 | |
| I0408 08:20:42.749770 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.117647 | |
| I0408 08:20:42.749788 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 08:20:42.749802 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.392157 | |
| I0408 08:20:42.749820 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.58254 (* 0.3 = 0.774762 loss) | |
| I0408 08:20:42.749835 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.792316 (* 0.3 = 0.237695 loss) | |
| I0408 08:20:42.749848 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.294118 | |
| I0408 08:20:42.749861 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 08:20:42.749874 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.647059 | |
| I0408 08:20:42.749891 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.31662 (* 0.3 = 0.694987 loss) | |
| I0408 08:20:42.749904 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.71464 (* 0.3 = 0.214392 loss) | |
| I0408 08:20:42.749917 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.607843 | |
| I0408 08:20:42.749930 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 08:20:42.749943 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.803922 | |
| I0408 08:20:42.749956 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.42669 (* 1 = 1.42669 loss) | |
| I0408 08:20:42.749971 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.441891 (* 1 = 0.441891 loss) | |
| I0408 08:20:42.749984 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 08:20:42.749997 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0416125 | |
| I0408 08:20:42.750012 3443 sgd_solver.cpp:106] Iteration 45000, lr = 0.00935714 | |
| I0408 08:26:16.039364 3443 solver.cpp:229] Iteration 45500, loss = 4.14317 | |
| I0408 08:26:16.039577 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.270833 | |
| I0408 08:26:16.039599 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 08:26:16.039614 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.375 | |
| I0408 08:26:16.039633 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.9851 (* 0.3 = 0.895529 loss) | |
| I0408 08:26:16.039647 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.880025 (* 0.3 = 0.264008 loss) | |
| I0408 08:26:16.039661 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.3125 | |
| I0408 08:26:16.039674 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 08:26:16.039687 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.458333 | |
| I0408 08:26:16.039703 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.53239 (* 0.3 = 0.759717 loss) | |
| I0408 08:26:16.039718 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.804404 (* 0.3 = 0.241321 loss) | |
| I0408 08:26:16.039731 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.479167 | |
| I0408 08:26:16.039743 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.840909 | |
| I0408 08:26:16.039757 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.708333 | |
| I0408 08:26:16.039772 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.82212 (* 1 = 1.82212 loss) | |
| I0408 08:26:16.039785 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.545502 (* 1 = 0.545502 loss) | |
| I0408 08:26:16.039798 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:26:16.039810 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0557219 | |
| I0408 08:26:16.039826 3443 sgd_solver.cpp:106] Iteration 45500, lr = 0.00935 | |
| I0408 08:31:50.326874 3443 solver.cpp:229] Iteration 46000, loss = 4.16537 | |
| I0408 08:31:50.327046 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.212766 | |
| I0408 08:31:50.327069 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 08:31:50.327082 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.361702 | |
| I0408 08:31:50.327100 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.73033 (* 0.3 = 0.819098 loss) | |
| I0408 08:31:50.327116 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.775577 (* 0.3 = 0.232673 loss) | |
| I0408 08:31:50.327128 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.255319 | |
| I0408 08:31:50.327142 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 08:31:50.327153 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.595745 | |
| I0408 08:31:50.327168 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.42142 (* 0.3 = 0.726427 loss) | |
| I0408 08:31:50.327183 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.681215 (* 0.3 = 0.204364 loss) | |
| I0408 08:31:50.327196 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.510638 | |
| I0408 08:31:50.327209 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 08:31:50.327221 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.87234 | |
| I0408 08:31:50.327236 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.7496 (* 1 = 1.7496 loss) | |
| I0408 08:31:50.327250 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.510245 (* 1 = 0.510245 loss) | |
| I0408 08:31:50.327263 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:31:50.327276 3443 solver.cpp:245] Train net output #16: total_confidence = 0.037237 | |
| I0408 08:31:50.327291 3443 sgd_solver.cpp:106] Iteration 46000, lr = 0.00934286 | |
| I0408 08:37:23.921041 3443 solver.cpp:229] Iteration 46500, loss = 4.12648 | |
| I0408 08:37:23.921207 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.232558 | |
| I0408 08:37:23.921229 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 08:37:23.921242 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.488372 | |
| I0408 08:37:23.921260 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.34259 (* 0.3 = 0.702776 loss) | |
| I0408 08:37:23.921277 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.642685 (* 0.3 = 0.192806 loss) | |
| I0408 08:37:23.921290 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.44186 | |
| I0408 08:37:23.921303 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0408 08:37:23.921316 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.767442 | |
| I0408 08:37:23.921331 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.82075 (* 0.3 = 0.546224 loss) | |
| I0408 08:37:23.921345 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.500723 (* 0.3 = 0.150217 loss) | |
| I0408 08:37:23.921358 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.72093 | |
| I0408 08:37:23.921371 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0408 08:37:23.921383 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.883721 | |
| I0408 08:37:23.921398 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.990515 (* 1 = 0.990515 loss) | |
| I0408 08:37:23.921412 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.296504 (* 1 = 0.296504 loss) | |
| I0408 08:37:23.921425 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 08:37:23.921438 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0528966 | |
| I0408 08:37:23.921453 3443 sgd_solver.cpp:106] Iteration 46500, lr = 0.00933571 | |
| I0408 08:42:58.185261 3443 solver.cpp:229] Iteration 47000, loss = 4.08458 | |
| I0408 08:42:58.185442 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.130435 | |
| I0408 08:42:58.185462 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 08:42:58.185477 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.369565 | |
| I0408 08:42:58.185495 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.90335 (* 0.3 = 0.871004 loss) | |
| I0408 08:42:58.185513 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.868988 (* 0.3 = 0.260696 loss) | |
| I0408 08:42:58.185525 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.23913 | |
| I0408 08:42:58.185539 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 08:42:58.185551 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.478261 | |
| I0408 08:42:58.185566 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.54151 (* 0.3 = 0.762454 loss) | |
| I0408 08:42:58.185580 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.762412 (* 0.3 = 0.228724 loss) | |
| I0408 08:42:58.185593 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.413043 | |
| I0408 08:42:58.185606 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.823864 | |
| I0408 08:42:58.185618 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.717391 | |
| I0408 08:42:58.185633 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.69967 (* 1 = 1.69967 loss) | |
| I0408 08:42:58.185648 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.545159 (* 1 = 0.545159 loss) | |
| I0408 08:42:58.185662 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:42:58.185673 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0416656 | |
| I0408 08:42:58.185688 3443 sgd_solver.cpp:106] Iteration 47000, lr = 0.00932857 | |
| I0408 08:48:31.634920 3443 solver.cpp:229] Iteration 47500, loss = 4.0875 | |
| I0408 08:48:31.635064 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.148936 | |
| I0408 08:48:31.635087 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 08:48:31.635100 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.340426 | |
| I0408 08:48:31.635118 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.21337 (* 0.3 = 0.964011 loss) | |
| I0408 08:48:31.635133 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.00573 (* 0.3 = 0.30172 loss) | |
| I0408 08:48:31.635148 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.234043 | |
| I0408 08:48:31.635160 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 08:48:31.635172 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.361702 | |
| I0408 08:48:31.635187 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.77413 (* 0.3 = 0.832239 loss) | |
| I0408 08:48:31.635202 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.815685 (* 0.3 = 0.244705 loss) | |
| I0408 08:48:31.635215 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.340426 | |
| I0408 08:48:31.635227 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.806818 | |
| I0408 08:48:31.635241 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.680851 | |
| I0408 08:48:31.635254 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.19916 (* 1 = 2.19916 loss) | |
| I0408 08:48:31.635269 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.655697 (* 1 = 0.655697 loss) | |
| I0408 08:48:31.635282 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 08:48:31.635294 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0168393 | |
| I0408 08:48:31.635310 3443 sgd_solver.cpp:106] Iteration 47500, lr = 0.00932143 | |
| I0408 08:54:05.018049 3443 solver.cpp:229] Iteration 48000, loss = 4.01293 | |
| I0408 08:54:05.018231 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.130435 | |
| I0408 08:54:05.018252 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 08:54:05.018266 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.478261 | |
| I0408 08:54:05.018283 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.60688 (* 0.3 = 0.782065 loss) | |
| I0408 08:54:05.018299 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.738723 (* 0.3 = 0.221617 loss) | |
| I0408 08:54:05.018312 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.326087 | |
| I0408 08:54:05.018326 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 08:54:05.018337 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.565217 | |
| I0408 08:54:05.018352 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.23764 (* 0.3 = 0.671293 loss) | |
| I0408 08:54:05.018368 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.652562 (* 0.3 = 0.195769 loss) | |
| I0408 08:54:05.018380 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.456522 | |
| I0408 08:54:05.018393 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.846591 | |
| I0408 08:54:05.018405 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.76087 | |
| I0408 08:54:05.018419 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.72949 (* 1 = 1.72949 loss) | |
| I0408 08:54:05.018435 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.496588 (* 1 = 0.496588 loss) | |
| I0408 08:54:05.018447 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 08:54:05.018458 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0332214 | |
| I0408 08:54:05.018474 3443 sgd_solver.cpp:106] Iteration 48000, lr = 0.00931429 | |
| I0408 08:59:38.758241 3443 solver.cpp:229] Iteration 48500, loss = 3.9872 | |
| I0408 08:59:38.758395 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.113636 | |
| I0408 08:59:38.758416 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 08:59:38.758430 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.431818 | |
| I0408 08:59:38.758447 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.86639 (* 0.3 = 0.859918 loss) | |
| I0408 08:59:38.758463 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.885945 (* 0.3 = 0.265783 loss) | |
| I0408 08:59:38.758476 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.227273 | |
| I0408 08:59:38.758489 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.761364 | |
| I0408 08:59:38.758502 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.5 | |
| I0408 08:59:38.758517 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.61334 (* 0.3 = 0.784002 loss) | |
| I0408 08:59:38.758533 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.895284 (* 0.3 = 0.268585 loss) | |
| I0408 08:59:38.758544 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.545455 | |
| I0408 08:59:38.758558 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.863636 | |
| I0408 08:59:38.758569 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.681818 | |
| I0408 08:59:38.758584 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.7929 (* 1 = 1.7929 loss) | |
| I0408 08:59:38.758599 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.580862 (* 1 = 0.580862 loss) | |
| I0408 08:59:38.758611 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 08:59:38.758623 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0750654 | |
| I0408 08:59:38.758638 3443 sgd_solver.cpp:106] Iteration 48500, lr = 0.00930714 | |
| I0408 09:05:12.239279 3443 solver.cpp:229] Iteration 49000, loss = 4.04302 | |
| I0408 09:05:12.239449 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.183673 | |
| I0408 09:05:12.239470 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 09:05:12.239485 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.367347 | |
| I0408 09:05:12.239500 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.74593 (* 0.3 = 0.823778 loss) | |
| I0408 09:05:12.239516 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.834283 (* 0.3 = 0.250285 loss) | |
| I0408 09:05:12.239529 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.285714 | |
| I0408 09:05:12.239542 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.772727 | |
| I0408 09:05:12.239554 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.55102 | |
| I0408 09:05:12.239569 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.3752 (* 0.3 = 0.71256 loss) | |
| I0408 09:05:12.239584 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.77604 (* 0.3 = 0.232812 loss) | |
| I0408 09:05:12.239600 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.428571 | |
| I0408 09:05:12.239614 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.829545 | |
| I0408 09:05:12.239626 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.653061 | |
| I0408 09:05:12.239641 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.64359 (* 1 = 1.64359 loss) | |
| I0408 09:05:12.239656 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.488295 (* 1 = 0.488295 loss) | |
| I0408 09:05:12.239668 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 09:05:12.239681 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0908568 | |
| I0408 09:05:12.239697 3443 sgd_solver.cpp:106] Iteration 49000, lr = 0.0093 | |
| I0408 09:10:45.819412 3443 solver.cpp:229] Iteration 49500, loss = 3.94229 | |
| I0408 09:10:45.819550 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.175 | |
| I0408 09:10:45.819571 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 09:10:45.819586 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.325 | |
| I0408 09:10:45.819602 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.74164 (* 0.3 = 0.822493 loss) | |
| I0408 09:10:45.819618 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.773376 (* 0.3 = 0.232013 loss) | |
| I0408 09:10:45.819631 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.275 | |
| I0408 09:10:45.819644 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 09:10:45.819656 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.5 | |
| I0408 09:10:45.819671 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.26481 (* 0.3 = 0.679442 loss) | |
| I0408 09:10:45.819686 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.61455 (* 0.3 = 0.184365 loss) | |
| I0408 09:10:45.819700 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.575 | |
| I0408 09:10:45.819711 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 09:10:45.819725 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.825 | |
| I0408 09:10:45.819739 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.40893 (* 1 = 1.40893 loss) | |
| I0408 09:10:45.819754 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.415714 (* 1 = 0.415714 loss) | |
| I0408 09:10:45.819767 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 09:10:45.819779 3443 solver.cpp:245] Train net output #16: total_confidence = 0.087181 | |
| I0408 09:10:45.819794 3443 sgd_solver.cpp:106] Iteration 49500, lr = 0.00929286 | |
| I0408 09:16:18.860211 3443 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_50000.caffemodel | |
| I0408 09:16:19.424011 3443 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_50000.solverstate | |
| I0408 09:16:19.682634 3443 solver.cpp:338] Iteration 50000, Testing net (#0) | |
| I0408 09:17:01.030030 3443 solver.cpp:393] Test loss: 3.50954 | |
| I0408 09:17:01.030125 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.190558 | |
| I0408 09:17:01.030144 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.796636 | |
| I0408 09:17:01.030159 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.443852 | |
| I0408 09:17:01.030175 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.80606 (* 0.3 = 0.841817 loss) | |
| I0408 09:17:01.030190 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.727638 (* 0.3 = 0.218291 loss) | |
| I0408 09:17:01.030203 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.430367 | |
| I0408 09:17:01.030215 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.850548 | |
| I0408 09:17:01.030227 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.730112 | |
| I0408 09:17:01.030241 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.90238 (* 0.3 = 0.570713 loss) | |
| I0408 09:17:01.030256 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.507911 (* 0.3 = 0.152373 loss) | |
| I0408 09:17:01.030268 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.617329 | |
| I0408 09:17:01.030280 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.895504 | |
| I0408 09:17:01.030292 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.848691 | |
| I0408 09:17:01.030306 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.35811 (* 1 = 1.35811 loss) | |
| I0408 09:17:01.030320 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.368237 (* 1 = 0.368237 loss) | |
| I0408 09:17:01.030333 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.06 | |
| I0408 09:17:01.030344 3443 solver.cpp:406] Test net output #16: total_confidence = 0.0745652 | |
| I0408 09:17:01.404201 3443 solver.cpp:229] Iteration 50000, loss = 3.93725 | |
| I0408 09:17:01.404254 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.348837 | |
| I0408 09:17:01.404273 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 09:17:01.404286 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.488372 | |
| I0408 09:17:01.404302 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.44191 (* 0.3 = 0.732572 loss) | |
| I0408 09:17:01.404317 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.775296 (* 0.3 = 0.232589 loss) | |
| I0408 09:17:01.404330 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.348837 | |
| I0408 09:17:01.404343 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 09:17:01.404356 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.697674 | |
| I0408 09:17:01.404371 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.05708 (* 0.3 = 0.617125 loss) | |
| I0408 09:17:01.404386 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.709426 (* 0.3 = 0.212828 loss) | |
| I0408 09:17:01.404399 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.651163 | |
| I0408 09:17:01.404412 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 09:17:01.404424 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.860465 | |
| I0408 09:17:01.404441 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.23018 (* 1 = 1.23018 loss) | |
| I0408 09:17:01.404458 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.499838 (* 1 = 0.499838 loss) | |
| I0408 09:17:01.404470 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 09:17:01.404482 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0903435 | |
| I0408 09:17:01.404497 3443 sgd_solver.cpp:106] Iteration 50000, lr = 0.00928571 | |
| I0408 09:22:34.903231 3443 solver.cpp:229] Iteration 50500, loss = 3.87351 | |
| I0408 09:22:34.903419 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.309524 | |
| I0408 09:22:34.903440 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 09:22:34.903455 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.547619 | |
| I0408 09:22:34.903472 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.39885 (* 0.3 = 0.719654 loss) | |
| I0408 09:22:34.903487 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.730287 (* 0.3 = 0.219086 loss) | |
| I0408 09:22:34.903501 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.380952 | |
| I0408 09:22:34.903513 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0408 09:22:34.903527 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.619048 | |
| I0408 09:22:34.903540 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.15242 (* 0.3 = 0.645726 loss) | |
| I0408 09:22:34.903555 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.648724 (* 0.3 = 0.194617 loss) | |
| I0408 09:22:34.903568 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.619048 | |
| I0408 09:22:34.903581 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0408 09:22:34.903594 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.785714 | |
| I0408 09:22:34.903609 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.35736 (* 1 = 1.35736 loss) | |
| I0408 09:22:34.903623 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.410744 (* 1 = 0.410744 loss) | |
| I0408 09:22:34.903637 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 09:22:34.903650 3443 solver.cpp:245] Train net output #16: total_confidence = 0.101376 | |
| I0408 09:22:34.903666 3443 sgd_solver.cpp:106] Iteration 50500, lr = 0.00927857 | |
| I0408 09:28:08.306874 3443 solver.cpp:229] Iteration 51000, loss = 3.89955 | |
| I0408 09:28:08.307045 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.134615 | |
| I0408 09:28:08.307068 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0408 09:28:08.307082 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.365385 | |
| I0408 09:28:08.307101 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.78082 (* 0.3 = 0.834245 loss) | |
| I0408 09:28:08.307116 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.881471 (* 0.3 = 0.264441 loss) | |
| I0408 09:28:08.307129 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.346154 | |
| I0408 09:28:08.307142 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 09:28:08.307155 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.519231 | |
| I0408 09:28:08.307170 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.46918 (* 0.3 = 0.740754 loss) | |
| I0408 09:28:08.307185 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.750153 (* 0.3 = 0.225046 loss) | |
| I0408 09:28:08.307198 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.384615 | |
| I0408 09:28:08.307210 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.795455 | |
| I0408 09:28:08.307224 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.615385 | |
| I0408 09:28:08.307238 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.98868 (* 1 = 1.98868 loss) | |
| I0408 09:28:08.307253 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.670761 (* 1 = 0.670761 loss) | |
| I0408 09:28:08.307265 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 09:28:08.307278 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0280175 | |
| I0408 09:28:08.307293 3443 sgd_solver.cpp:106] Iteration 51000, lr = 0.00927143 | |
| I0408 09:33:42.101816 3443 solver.cpp:229] Iteration 51500, loss = 3.84542 | |
| I0408 09:33:42.102283 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.190476 | |
| I0408 09:33:42.102306 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 09:33:42.102321 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.380952 | |
| I0408 09:33:42.102340 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.25572 (* 0.3 = 0.976717 loss) | |
| I0408 09:33:42.102355 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.89719 (* 0.3 = 0.269157 loss) | |
| I0408 09:33:42.102368 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.285714 | |
| I0408 09:33:42.102382 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 09:33:42.102394 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.52381 | |
| I0408 09:33:42.102409 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.64087 (* 0.3 = 0.792262 loss) | |
| I0408 09:33:42.102424 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.752167 (* 0.3 = 0.22565 loss) | |
| I0408 09:33:42.102437 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.547619 | |
| I0408 09:33:42.102450 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.869318 | |
| I0408 09:33:42.102463 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.714286 | |
| I0408 09:33:42.102478 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.68187 (* 1 = 1.68187 loss) | |
| I0408 09:33:42.102493 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.448615 (* 1 = 0.448615 loss) | |
| I0408 09:33:42.102505 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 09:33:42.102517 3443 solver.cpp:245] Train net output #16: total_confidence = 0.16407 | |
| I0408 09:33:42.102533 3443 sgd_solver.cpp:106] Iteration 51500, lr = 0.00926429 | |
| I0408 09:39:15.541847 3443 solver.cpp:229] Iteration 52000, loss = 3.86542 | |
| I0408 09:39:15.541996 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.325 | |
| I0408 09:39:15.542016 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0408 09:39:15.542031 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.475 | |
| I0408 09:39:15.542048 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.54118 (* 0.3 = 0.762355 loss) | |
| I0408 09:39:15.542063 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.698118 (* 0.3 = 0.209436 loss) | |
| I0408 09:39:15.542076 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.425 | |
| I0408 09:39:15.542089 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 09:39:15.542103 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.725 | |
| I0408 09:39:15.542116 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.92787 (* 0.3 = 0.57836 loss) | |
| I0408 09:39:15.542131 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.570286 (* 0.3 = 0.171086 loss) | |
| I0408 09:39:15.542143 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.75 | |
| I0408 09:39:15.542156 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0408 09:39:15.542170 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.95 | |
| I0408 09:39:15.542183 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.93207 (* 1 = 0.93207 loss) | |
| I0408 09:39:15.542198 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.284772 (* 1 = 0.284772 loss) | |
| I0408 09:39:15.542212 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 09:39:15.542223 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0789691 | |
| I0408 09:39:15.542239 3443 sgd_solver.cpp:106] Iteration 52000, lr = 0.00925714 | |
| I0408 09:44:48.919039 3443 solver.cpp:229] Iteration 52500, loss = 3.80522 | |
| I0408 09:44:48.919539 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.24 | |
| I0408 09:44:48.919562 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 09:44:48.919577 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.44 | |
| I0408 09:44:48.919595 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.76324 (* 0.3 = 0.828973 loss) | |
| I0408 09:44:48.919610 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.823641 (* 0.3 = 0.247092 loss) | |
| I0408 09:44:48.919625 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.3 | |
| I0408 09:44:48.919636 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0408 09:44:48.919649 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.62 | |
| I0408 09:44:48.919664 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.21419 (* 0.3 = 0.664257 loss) | |
| I0408 09:44:48.919679 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.666676 (* 0.3 = 0.200003 loss) | |
| I0408 09:44:48.919692 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.56 | |
| I0408 09:44:48.919704 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 09:44:48.919718 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.82 | |
| I0408 09:44:48.919731 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.53753 (* 1 = 1.53753 loss) | |
| I0408 09:44:48.919747 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.485502 (* 1 = 0.485502 loss) | |
| I0408 09:44:48.919760 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 09:44:48.919772 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0392622 | |
| I0408 09:44:48.919787 3443 sgd_solver.cpp:106] Iteration 52500, lr = 0.00925 | |
| I0408 09:50:22.271160 3443 solver.cpp:229] Iteration 53000, loss = 3.86425 | |
| I0408 09:50:22.271280 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.229167 | |
| I0408 09:50:22.271311 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 09:50:22.271359 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.479167 | |
| I0408 09:50:22.271392 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.79969 (* 0.3 = 0.839907 loss) | |
| I0408 09:50:22.271422 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.81329 (* 0.3 = 0.243987 loss) | |
| I0408 09:50:22.271445 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.354167 | |
| I0408 09:50:22.271469 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 09:50:22.271492 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 09:50:22.271518 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.25262 (* 0.3 = 0.675785 loss) | |
| I0408 09:50:22.271544 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.663965 (* 0.3 = 0.199189 loss) | |
| I0408 09:50:22.271569 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.541667 | |
| I0408 09:50:22.271591 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.869318 | |
| I0408 09:50:22.271616 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.8125 | |
| I0408 09:50:22.271647 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.70103 (* 1 = 1.70103 loss) | |
| I0408 09:50:22.271677 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.487557 (* 1 = 0.487557 loss) | |
| I0408 09:50:22.271702 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 09:50:22.271723 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0639214 | |
| I0408 09:50:22.271749 3443 sgd_solver.cpp:106] Iteration 53000, lr = 0.00924286 | |
| I0408 09:55:55.908529 3443 solver.cpp:229] Iteration 53500, loss = 3.7541 | |
| I0408 09:55:55.908932 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170213 | |
| I0408 09:55:55.908954 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 09:55:55.908967 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.425532 | |
| I0408 09:55:55.908984 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.92748 (* 0.3 = 0.878245 loss) | |
| I0408 09:55:55.908999 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.832402 (* 0.3 = 0.249721 loss) | |
| I0408 09:55:55.909013 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.191489 | |
| I0408 09:55:55.909025 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 09:55:55.909037 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.404255 | |
| I0408 09:55:55.909051 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.62275 (* 0.3 = 0.786826 loss) | |
| I0408 09:55:55.909066 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.743678 (* 0.3 = 0.223103 loss) | |
| I0408 09:55:55.909078 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.638298 | |
| I0408 09:55:55.909091 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0408 09:55:55.909103 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.829787 | |
| I0408 09:55:55.909117 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.26441 (* 1 = 1.26441 loss) | |
| I0408 09:55:55.909132 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.367668 (* 1 = 0.367668 loss) | |
| I0408 09:55:55.909145 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 09:55:55.909157 3443 solver.cpp:245] Train net output #16: total_confidence = 0.038417 | |
| I0408 09:55:55.909173 3443 sgd_solver.cpp:106] Iteration 53500, lr = 0.00923571 | |
| I0408 10:01:29.365522 3443 solver.cpp:229] Iteration 54000, loss = 3.77631 | |
| I0408 10:01:29.365659 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170213 | |
| I0408 10:01:29.365680 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 10:01:29.365694 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.382979 | |
| I0408 10:01:29.365710 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.7261 (* 0.3 = 0.817831 loss) | |
| I0408 10:01:29.365725 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.797582 (* 0.3 = 0.239275 loss) | |
| I0408 10:01:29.365738 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.361702 | |
| I0408 10:01:29.365751 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 10:01:29.365763 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.574468 | |
| I0408 10:01:29.365777 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.52451 (* 0.3 = 0.757354 loss) | |
| I0408 10:01:29.365792 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.720904 (* 0.3 = 0.216271 loss) | |
| I0408 10:01:29.365804 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.574468 | |
| I0408 10:01:29.365818 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 10:01:29.365830 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.723404 | |
| I0408 10:01:29.365844 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.04732 (* 1 = 2.04732 loss) | |
| I0408 10:01:29.365859 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.563827 (* 1 = 0.563827 loss) | |
| I0408 10:01:29.365872 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 10:01:29.365885 3443 solver.cpp:245] Train net output #16: total_confidence = 0.119996 | |
| I0408 10:01:29.365908 3443 sgd_solver.cpp:106] Iteration 54000, lr = 0.00922857 | |
| I0408 10:07:02.687732 3443 solver.cpp:229] Iteration 54500, loss = 3.7368 | |
| I0408 10:07:02.688166 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.226415 | |
| I0408 10:07:02.688189 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 10:07:02.688202 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.433962 | |
| I0408 10:07:02.688220 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.56945 (* 0.3 = 0.770836 loss) | |
| I0408 10:07:02.688235 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.796454 (* 0.3 = 0.238936 loss) | |
| I0408 10:07:02.688247 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.339623 | |
| I0408 10:07:02.688261 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 10:07:02.688273 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.566038 | |
| I0408 10:07:02.688287 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.17992 (* 0.3 = 0.653976 loss) | |
| I0408 10:07:02.688302 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.683631 (* 0.3 = 0.205089 loss) | |
| I0408 10:07:02.688315 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.54717 | |
| I0408 10:07:02.688328 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 10:07:02.688340 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.754717 | |
| I0408 10:07:02.688356 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.45734 (* 1 = 1.45734 loss) | |
| I0408 10:07:02.688371 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.457991 (* 1 = 0.457991 loss) | |
| I0408 10:07:02.688383 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 10:07:02.688395 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0264492 | |
| I0408 10:07:02.688411 3443 sgd_solver.cpp:106] Iteration 54500, lr = 0.00922143 | |
| I0408 10:12:35.682016 3443 solver.cpp:338] Iteration 55000, Testing net (#0) | |
| I0408 10:13:16.585360 3443 solver.cpp:393] Test loss: 3.40197 | |
| I0408 10:13:16.585489 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.210996 | |
| I0408 10:13:16.585510 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.805409 | |
| I0408 10:13:16.585523 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.47487 | |
| I0408 10:13:16.585539 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.71706 (* 0.3 = 0.815118 loss) | |
| I0408 10:13:16.585554 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.684171 (* 0.3 = 0.205251 loss) | |
| I0408 10:13:16.585568 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.401732 | |
| I0408 10:13:16.585580 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.849866 | |
| I0408 10:13:16.585592 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.746683 | |
| I0408 10:13:16.585607 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.92677 (* 0.3 = 0.578031 loss) | |
| I0408 10:13:16.585620 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.489848 (* 0.3 = 0.146955 loss) | |
| I0408 10:13:16.585631 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.627243 | |
| I0408 10:13:16.585644 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.905775 | |
| I0408 10:13:16.585655 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.865385 | |
| I0408 10:13:16.585669 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.32215 (* 1 = 1.32215 loss) | |
| I0408 10:13:16.585683 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.334463 (* 1 = 0.334463 loss) | |
| I0408 10:13:16.585695 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.14 | |
| I0408 10:13:16.585707 3443 solver.cpp:406] Test net output #16: total_confidence = 0.172498 | |
| I0408 10:13:16.958045 3443 solver.cpp:229] Iteration 55000, loss = 3.72537 | |
| I0408 10:13:16.958096 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.315789 | |
| I0408 10:13:16.958115 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0408 10:13:16.958128 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.578947 | |
| I0408 10:13:16.958145 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.15683 (* 0.3 = 0.64705 loss) | |
| I0408 10:13:16.958160 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.550057 (* 0.3 = 0.165017 loss) | |
| I0408 10:13:16.958173 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0408 10:13:16.958186 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0408 10:13:16.958199 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.763158 | |
| I0408 10:13:16.958214 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.7039 (* 0.3 = 0.511171 loss) | |
| I0408 10:13:16.958228 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.465891 (* 0.3 = 0.139767 loss) | |
| I0408 10:13:16.958240 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.736842 | |
| I0408 10:13:16.958253 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0408 10:13:16.958266 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.947368 | |
| I0408 10:13:16.958279 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.741311 (* 1 = 0.741311 loss) | |
| I0408 10:13:16.958294 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.18732 (* 1 = 0.18732 loss) | |
| I0408 10:13:16.958307 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 10:13:16.958319 3443 solver.cpp:245] Train net output #16: total_confidence = 0.210917 | |
| I0408 10:13:16.958334 3443 sgd_solver.cpp:106] Iteration 55000, lr = 0.00921429 | |
| I0408 10:18:50.460870 3443 solver.cpp:229] Iteration 55500, loss = 3.7359 | |
| I0408 10:18:50.460973 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.195652 | |
| I0408 10:18:50.460993 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 10:18:50.461005 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.5 | |
| I0408 10:18:50.461022 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.56535 (* 0.3 = 0.769605 loss) | |
| I0408 10:18:50.461038 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.751759 (* 0.3 = 0.225528 loss) | |
| I0408 10:18:50.461051 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.456522 | |
| I0408 10:18:50.461064 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0408 10:18:50.461076 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.695652 | |
| I0408 10:18:50.461091 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.91438 (* 0.3 = 0.574315 loss) | |
| I0408 10:18:50.461105 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.551228 (* 0.3 = 0.165368 loss) | |
| I0408 10:18:50.461117 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.76087 | |
| I0408 10:18:50.461130 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0408 10:18:50.461143 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0408 10:18:50.461158 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.737849 (* 1 = 0.737849 loss) | |
| I0408 10:18:50.461170 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.268415 (* 1 = 0.268415 loss) | |
| I0408 10:18:50.461184 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 10:18:50.461195 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0579981 | |
| I0408 10:18:50.461210 3443 sgd_solver.cpp:106] Iteration 55500, lr = 0.00920714 | |
| I0408 10:24:23.996737 3443 solver.cpp:229] Iteration 56000, loss = 3.7612 | |
| I0408 10:24:23.997128 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.108696 | |
| I0408 10:24:23.997148 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0408 10:24:23.997164 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.195652 | |
| I0408 10:24:23.997180 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.18501 (* 0.3 = 0.955503 loss) | |
| I0408 10:24:23.997195 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.07651 (* 0.3 = 0.322953 loss) | |
| I0408 10:24:23.997208 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.173913 | |
| I0408 10:24:23.997221 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0408 10:24:23.997233 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.326087 | |
| I0408 10:24:23.997248 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.23806 (* 0.3 = 0.971418 loss) | |
| I0408 10:24:23.997262 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.05891 (* 0.3 = 0.317673 loss) | |
| I0408 10:24:23.997275 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.5 | |
| I0408 10:24:23.997288 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.8125 | |
| I0408 10:24:23.997300 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.608696 | |
| I0408 10:24:23.997315 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.13831 (* 1 = 2.13831 loss) | |
| I0408 10:24:23.997329 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.805548 (* 1 = 0.805548 loss) | |
| I0408 10:24:23.997342 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 10:24:23.997354 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0402396 | |
| I0408 10:24:23.997370 3443 sgd_solver.cpp:106] Iteration 56000, lr = 0.0092 | |
| I0408 10:29:57.623093 3443 solver.cpp:229] Iteration 56500, loss = 3.70488 | |
| I0408 10:29:57.623283 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.254902 | |
| I0408 10:29:57.623306 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 10:29:57.623319 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.529412 | |
| I0408 10:29:57.623337 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.58184 (* 0.3 = 0.774551 loss) | |
| I0408 10:29:57.623353 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.772729 (* 0.3 = 0.231819 loss) | |
| I0408 10:29:57.623365 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.294118 | |
| I0408 10:29:57.623378 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 10:29:57.623391 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.607843 | |
| I0408 10:29:57.623421 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.3651 (* 0.3 = 0.709531 loss) | |
| I0408 10:29:57.623437 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.717205 (* 0.3 = 0.215161 loss) | |
| I0408 10:29:57.623450 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.568627 | |
| I0408 10:29:57.623463 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.869318 | |
| I0408 10:29:57.623476 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.764706 | |
| I0408 10:29:57.623491 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.44408 (* 1 = 1.44408 loss) | |
| I0408 10:29:57.623505 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.438478 (* 1 = 0.438478 loss) | |
| I0408 10:29:57.623518 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 10:29:57.623530 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0525542 | |
| I0408 10:29:57.623546 3443 sgd_solver.cpp:106] Iteration 56500, lr = 0.00919286 | |
| I0408 10:35:31.390022 3443 solver.cpp:229] Iteration 57000, loss = 3.66924 | |
| I0408 10:35:31.390405 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.183673 | |
| I0408 10:35:31.390427 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 10:35:31.390441 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.346939 | |
| I0408 10:35:31.390460 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.77391 (* 0.3 = 0.832174 loss) | |
| I0408 10:35:31.390475 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.836753 (* 0.3 = 0.251026 loss) | |
| I0408 10:35:31.390487 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.285714 | |
| I0408 10:35:31.390501 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 10:35:31.390513 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.44898 | |
| I0408 10:35:31.390527 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.50232 (* 0.3 = 0.750696 loss) | |
| I0408 10:35:31.390542 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.754554 (* 0.3 = 0.226366 loss) | |
| I0408 10:35:31.390555 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.489796 | |
| I0408 10:35:31.390568 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.835227 | |
| I0408 10:35:31.390581 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.795918 | |
| I0408 10:35:31.390595 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.78511 (* 1 = 1.78511 loss) | |
| I0408 10:35:31.390610 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.550956 (* 1 = 0.550956 loss) | |
| I0408 10:35:31.390624 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 10:35:31.390635 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0448623 | |
| I0408 10:35:31.390650 3443 sgd_solver.cpp:106] Iteration 57000, lr = 0.00918571 | |
| I0408 10:41:05.519536 3443 solver.cpp:229] Iteration 57500, loss = 3.67591 | |
| I0408 10:41:05.519701 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.340426 | |
| I0408 10:41:05.519721 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 10:41:05.519737 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.680851 | |
| I0408 10:41:05.519754 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.38514 (* 0.3 = 0.715542 loss) | |
| I0408 10:41:05.519770 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.745789 (* 0.3 = 0.223737 loss) | |
| I0408 10:41:05.519783 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.425532 | |
| I0408 10:41:05.519796 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 10:41:05.519809 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.617021 | |
| I0408 10:41:05.519824 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.11699 (* 0.3 = 0.635097 loss) | |
| I0408 10:41:05.519839 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.672115 (* 0.3 = 0.201635 loss) | |
| I0408 10:41:05.519851 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.744681 | |
| I0408 10:41:05.519865 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 10:41:05.519877 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0408 10:41:05.519892 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.844032 (* 1 = 0.844032 loss) | |
| I0408 10:41:05.519907 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.273503 (* 1 = 0.273503 loss) | |
| I0408 10:41:05.519923 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 10:41:05.519937 3443 solver.cpp:245] Train net output #16: total_confidence = 0.131857 | |
| I0408 10:41:05.519953 3443 sgd_solver.cpp:106] Iteration 57500, lr = 0.00917857 | |
| I0408 10:46:39.566627 3443 solver.cpp:229] Iteration 58000, loss = 3.64727 | |
| I0408 10:46:39.567076 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.26 | |
| I0408 10:46:39.567101 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 10:46:39.567116 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.46 | |
| I0408 10:46:39.567132 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.40367 (* 0.3 = 0.721102 loss) | |
| I0408 10:46:39.567148 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.719607 (* 0.3 = 0.215882 loss) | |
| I0408 10:46:39.567162 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.28 | |
| I0408 10:46:39.567175 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 10:46:39.567188 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.66 | |
| I0408 10:46:39.567203 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.06164 (* 0.3 = 0.618493 loss) | |
| I0408 10:46:39.567219 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.629814 (* 0.3 = 0.188944 loss) | |
| I0408 10:46:39.567235 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.72 | |
| I0408 10:46:39.567263 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 10:46:39.567287 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9 | |
| I0408 10:46:39.567304 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.876963 (* 1 = 0.876963 loss) | |
| I0408 10:46:39.567342 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.354774 (* 1 = 0.354774 loss) | |
| I0408 10:46:39.567358 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 10:46:39.567369 3443 solver.cpp:245] Train net output #16: total_confidence = 0.073232 | |
| I0408 10:46:39.567385 3443 sgd_solver.cpp:106] Iteration 58000, lr = 0.00917143 | |
| I0408 10:52:12.899822 3443 solver.cpp:229] Iteration 58500, loss = 3.60253 | |
| I0408 10:52:12.900231 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170213 | |
| I0408 10:52:12.900254 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 10:52:12.900267 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.446809 | |
| I0408 10:52:12.900286 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.80731 (* 0.3 = 0.842193 loss) | |
| I0408 10:52:12.900302 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.805183 (* 0.3 = 0.241555 loss) | |
| I0408 10:52:12.900315 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.234043 | |
| I0408 10:52:12.900328 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 10:52:12.900341 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.489362 | |
| I0408 10:52:12.900357 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.435 (* 0.3 = 0.7305 loss) | |
| I0408 10:52:12.900372 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.701563 (* 0.3 = 0.210469 loss) | |
| I0408 10:52:12.900384 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.510638 | |
| I0408 10:52:12.900398 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.852273 | |
| I0408 10:52:12.900410 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.829787 | |
| I0408 10:52:12.900425 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.6553 (* 1 = 1.6553 loss) | |
| I0408 10:52:12.900439 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.506835 (* 1 = 0.506835 loss) | |
| I0408 10:52:12.900452 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 10:52:12.900465 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0403508 | |
| I0408 10:52:12.900480 3443 sgd_solver.cpp:106] Iteration 58500, lr = 0.00916429 | |
| I0408 10:57:46.536463 3443 solver.cpp:229] Iteration 59000, loss = 3.61423 | |
| I0408 10:57:46.536685 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.155556 | |
| I0408 10:57:46.536707 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 10:57:46.536721 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.4 | |
| I0408 10:57:46.536739 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.69887 (* 0.3 = 0.80966 loss) | |
| I0408 10:57:46.536754 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.75954 (* 0.3 = 0.227862 loss) | |
| I0408 10:57:46.536767 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.355556 | |
| I0408 10:57:46.536782 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 10:57:46.536793 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.622222 | |
| I0408 10:57:46.536808 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.10597 (* 0.3 = 0.631791 loss) | |
| I0408 10:57:46.536823 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.617394 (* 0.3 = 0.185218 loss) | |
| I0408 10:57:46.536836 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.733333 | |
| I0408 10:57:46.536849 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0408 10:57:46.536861 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.911111 | |
| I0408 10:57:46.536875 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.93498 (* 1 = 0.93498 loss) | |
| I0408 10:57:46.536890 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.252203 (* 1 = 0.252203 loss) | |
| I0408 10:57:46.536903 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 10:57:46.536916 3443 solver.cpp:245] Train net output #16: total_confidence = 0.08483 | |
| I0408 10:57:46.536934 3443 sgd_solver.cpp:106] Iteration 59000, lr = 0.00915714 | |
| I0408 11:03:19.916813 3443 solver.cpp:229] Iteration 59500, loss = 3.55155 | |
| I0408 11:03:19.917109 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.155556 | |
| I0408 11:03:19.917145 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 11:03:19.917165 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.4 | |
| I0408 11:03:19.917183 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.80527 (* 0.3 = 0.84158 loss) | |
| I0408 11:03:19.917201 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.78559 (* 0.3 = 0.235677 loss) | |
| I0408 11:03:19.917213 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.266667 | |
| I0408 11:03:19.917227 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 11:03:19.917239 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.577778 | |
| I0408 11:03:19.917253 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.30948 (* 0.3 = 0.692843 loss) | |
| I0408 11:03:19.917268 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.625826 (* 0.3 = 0.187748 loss) | |
| I0408 11:03:19.917281 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.533333 | |
| I0408 11:03:19.917294 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0408 11:03:19.917306 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.733333 | |
| I0408 11:03:19.917320 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.63747 (* 1 = 1.63747 loss) | |
| I0408 11:03:19.917335 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.438247 (* 1 = 0.438247 loss) | |
| I0408 11:03:19.917348 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 11:03:19.917361 3443 solver.cpp:245] Train net output #16: total_confidence = 0.125727 | |
| I0408 11:03:19.917376 3443 sgd_solver.cpp:106] Iteration 59500, lr = 0.00915 | |
| I0408 11:08:52.920253 3443 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_60000.caffemodel | |
| I0408 11:08:53.530155 3443 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_60000.solverstate | |
| I0408 11:08:53.793843 3443 solver.cpp:338] Iteration 60000, Testing net (#0) | |
| I0408 11:09:34.996886 3443 solver.cpp:393] Test loss: 3.48418 | |
| I0408 11:09:34.996976 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.187125 | |
| I0408 11:09:34.996995 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.799408 | |
| I0408 11:09:34.997009 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.445353 | |
| I0408 11:09:34.997026 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.88373 (* 0.3 = 0.86512 loss) | |
| I0408 11:09:34.997042 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.72653 (* 0.3 = 0.217959 loss) | |
| I0408 11:09:34.997054 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.437334 | |
| I0408 11:09:34.997066 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.856775 | |
| I0408 11:09:34.997078 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.759493 | |
| I0408 11:09:34.997092 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.87746 (* 0.3 = 0.563239 loss) | |
| I0408 11:09:34.997107 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.486616 (* 0.3 = 0.145985 loss) | |
| I0408 11:09:34.997118 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.633475 | |
| I0408 11:09:34.997131 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.905184 | |
| I0408 11:09:34.997143 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.859517 | |
| I0408 11:09:34.997158 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.34482 (* 1 = 1.34482 loss) | |
| I0408 11:09:34.997170 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.347059 (* 1 = 0.347059 loss) | |
| I0408 11:09:34.997184 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.166 | |
| I0408 11:09:34.997195 3443 solver.cpp:406] Test net output #16: total_confidence = 0.163634 | |
| I0408 11:09:35.371390 3443 solver.cpp:229] Iteration 60000, loss = 3.52198 | |
| I0408 11:09:35.371448 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.183673 | |
| I0408 11:09:35.371467 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 11:09:35.371481 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.387755 | |
| I0408 11:09:35.371498 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.52143 (* 0.3 = 0.756429 loss) | |
| I0408 11:09:35.371515 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.748841 (* 0.3 = 0.224652 loss) | |
| I0408 11:09:35.371532 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.22449 | |
| I0408 11:09:35.371546 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 11:09:35.371558 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.714286 | |
| I0408 11:09:35.371573 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.0764 (* 0.3 = 0.622919 loss) | |
| I0408 11:09:35.371588 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.612169 (* 0.3 = 0.183651 loss) | |
| I0408 11:09:35.371601 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.612245 | |
| I0408 11:09:35.371614 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 11:09:35.371626 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.857143 | |
| I0408 11:09:35.371641 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.2685 (* 1 = 1.2685 loss) | |
| I0408 11:09:35.371655 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.368635 (* 1 = 0.368635 loss) | |
| I0408 11:09:35.371668 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 11:09:35.371681 3443 solver.cpp:245] Train net output #16: total_confidence = 0.152606 | |
| I0408 11:09:35.371696 3443 sgd_solver.cpp:106] Iteration 60000, lr = 0.00914286 | |
| I0408 11:15:08.693100 3443 solver.cpp:229] Iteration 60500, loss = 3.52712 | |
| I0408 11:15:08.693475 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.297872 | |
| I0408 11:15:08.693497 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 11:15:08.693511 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.531915 | |
| I0408 11:15:08.693527 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.35567 (* 0.3 = 0.706702 loss) | |
| I0408 11:15:08.693542 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.689462 (* 0.3 = 0.206839 loss) | |
| I0408 11:15:08.693555 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.255319 | |
| I0408 11:15:08.693568 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 11:15:08.693581 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.638298 | |
| I0408 11:15:08.693594 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.15271 (* 0.3 = 0.645813 loss) | |
| I0408 11:15:08.693609 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.597349 (* 0.3 = 0.179205 loss) | |
| I0408 11:15:08.693621 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.574468 | |
| I0408 11:15:08.693634 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0408 11:15:08.693646 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.808511 | |
| I0408 11:15:08.693660 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.29279 (* 1 = 1.29279 loss) | |
| I0408 11:15:08.693675 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.391341 (* 1 = 0.391341 loss) | |
| I0408 11:15:08.693686 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 11:15:08.693698 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0946371 | |
| I0408 11:15:08.693712 3443 sgd_solver.cpp:106] Iteration 60500, lr = 0.00913571 | |
| I0408 11:20:42.074808 3443 solver.cpp:229] Iteration 61000, loss = 3.50445 | |
| I0408 11:20:42.074913 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.295455 | |
| I0408 11:20:42.074931 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 11:20:42.074944 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.477273 | |
| I0408 11:20:42.074961 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.51246 (* 0.3 = 0.753737 loss) | |
| I0408 11:20:42.074976 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.746655 (* 0.3 = 0.223996 loss) | |
| I0408 11:20:42.074991 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.386364 | |
| I0408 11:20:42.075004 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 11:20:42.075017 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.636364 | |
| I0408 11:20:42.075031 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.17316 (* 0.3 = 0.651947 loss) | |
| I0408 11:20:42.075045 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.627462 (* 0.3 = 0.188239 loss) | |
| I0408 11:20:42.075058 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.659091 | |
| I0408 11:20:42.075072 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0408 11:20:42.075083 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.909091 | |
| I0408 11:20:42.075098 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.07746 (* 1 = 1.07746 loss) | |
| I0408 11:20:42.075112 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.327231 (* 1 = 0.327231 loss) | |
| I0408 11:20:42.075125 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 11:20:42.075137 3443 solver.cpp:245] Train net output #16: total_confidence = 0.119539 | |
| I0408 11:20:42.075151 3443 sgd_solver.cpp:106] Iteration 61000, lr = 0.00912857 | |
| I0408 11:26:15.458230 3443 solver.cpp:229] Iteration 61500, loss = 3.49769 | |
| I0408 11:26:15.458529 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.377778 | |
| I0408 11:26:15.458554 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0408 11:26:15.458566 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.555556 | |
| I0408 11:26:15.458583 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.37871 (* 0.3 = 0.713613 loss) | |
| I0408 11:26:15.458598 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.667191 (* 0.3 = 0.200157 loss) | |
| I0408 11:26:15.458611 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4 | |
| I0408 11:26:15.458624 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 11:26:15.458636 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.733333 | |
| I0408 11:26:15.458650 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.04564 (* 0.3 = 0.613692 loss) | |
| I0408 11:26:15.458664 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.579142 (* 0.3 = 0.173743 loss) | |
| I0408 11:26:15.458676 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.755556 | |
| I0408 11:26:15.458689 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0408 11:26:15.458701 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.911111 | |
| I0408 11:26:15.458715 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.845799 (* 1 = 0.845799 loss) | |
| I0408 11:26:15.458729 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.255806 (* 1 = 0.255806 loss) | |
| I0408 11:26:15.458741 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 11:26:15.458755 3443 solver.cpp:245] Train net output #16: total_confidence = 0.173829 | |
| I0408 11:26:15.458767 3443 sgd_solver.cpp:106] Iteration 61500, lr = 0.00912143 | |
| I0408 11:31:48.849228 3443 solver.cpp:229] Iteration 62000, loss = 3.49207 | |
| I0408 11:31:48.849354 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.217391 | |
| I0408 11:31:48.849373 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 11:31:48.849386 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.543478 | |
| I0408 11:31:48.849403 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.58267 (* 0.3 = 0.7748 loss) | |
| I0408 11:31:48.849418 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.7413 (* 0.3 = 0.22239 loss) | |
| I0408 11:31:48.849431 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.369565 | |
| I0408 11:31:48.849443 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 11:31:48.849457 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.608696 | |
| I0408 11:31:48.849470 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.21541 (* 0.3 = 0.664622 loss) | |
| I0408 11:31:48.849484 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.647358 (* 0.3 = 0.194207 loss) | |
| I0408 11:31:48.849496 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.673913 | |
| I0408 11:31:48.849509 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0408 11:31:48.849520 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.891304 | |
| I0408 11:31:48.849535 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.11836 (* 1 = 1.11836 loss) | |
| I0408 11:31:48.849550 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.361928 (* 1 = 0.361928 loss) | |
| I0408 11:31:48.849562 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 11:31:48.849575 3443 solver.cpp:245] Train net output #16: total_confidence = 0.139846 | |
| I0408 11:31:48.849589 3443 sgd_solver.cpp:106] Iteration 62000, lr = 0.00911429 | |
| I0408 11:37:22.226192 3443 solver.cpp:229] Iteration 62500, loss = 3.44632 | |
| I0408 11:37:22.226552 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.304348 | |
| I0408 11:37:22.226572 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 11:37:22.226585 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.630435 | |
| I0408 11:37:22.226603 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.25842 (* 0.3 = 0.677527 loss) | |
| I0408 11:37:22.226618 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.653495 (* 0.3 = 0.196049 loss) | |
| I0408 11:37:22.226630 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.434783 | |
| I0408 11:37:22.226642 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0408 11:37:22.226655 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.695652 | |
| I0408 11:37:22.226668 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.87739 (* 0.3 = 0.563217 loss) | |
| I0408 11:37:22.226683 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.557116 (* 0.3 = 0.167135 loss) | |
| I0408 11:37:22.226696 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.717391 | |
| I0408 11:37:22.226707 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0408 11:37:22.226719 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0408 11:37:22.226734 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.975509 (* 1 = 0.975509 loss) | |
| I0408 11:37:22.226748 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.308414 (* 1 = 0.308414 loss) | |
| I0408 11:37:22.226760 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 11:37:22.226773 3443 solver.cpp:245] Train net output #16: total_confidence = 0.177911 | |
| I0408 11:37:22.226786 3443 sgd_solver.cpp:106] Iteration 62500, lr = 0.00910714 | |
| I0408 11:42:55.615823 3443 solver.cpp:229] Iteration 63000, loss = 3.4143 | |
| I0408 11:42:55.616144 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.282609 | |
| I0408 11:42:55.616166 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 11:42:55.616180 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.543478 | |
| I0408 11:42:55.616197 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.55631 (* 0.3 = 0.766894 loss) | |
| I0408 11:42:55.616214 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.76566 (* 0.3 = 0.229698 loss) | |
| I0408 11:42:55.616226 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.413043 | |
| I0408 11:42:55.616240 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0408 11:42:55.616253 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.652174 | |
| I0408 11:42:55.616267 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.17226 (* 0.3 = 0.651679 loss) | |
| I0408 11:42:55.616281 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.682501 (* 0.3 = 0.20475 loss) | |
| I0408 11:42:55.616294 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.608696 | |
| I0408 11:42:55.616307 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0408 11:42:55.616318 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.804348 | |
| I0408 11:42:55.616333 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.66186 (* 1 = 1.66186 loss) | |
| I0408 11:42:55.616348 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.470194 (* 1 = 0.470194 loss) | |
| I0408 11:42:55.616359 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 11:42:55.616371 3443 solver.cpp:245] Train net output #16: total_confidence = 0.143958 | |
| I0408 11:42:55.616386 3443 sgd_solver.cpp:106] Iteration 63000, lr = 0.0091 | |
| I0408 11:48:28.999984 3443 solver.cpp:229] Iteration 63500, loss = 3.43353 | |
| I0408 11:48:29.000146 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.295455 | |
| I0408 11:48:29.000167 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 11:48:29.000181 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.568182 | |
| I0408 11:48:29.000200 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.38163 (* 0.3 = 0.71449 loss) | |
| I0408 11:48:29.000214 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.64411 (* 0.3 = 0.193233 loss) | |
| I0408 11:48:29.000227 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.454545 | |
| I0408 11:48:29.000239 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0408 11:48:29.000252 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.75 | |
| I0408 11:48:29.000265 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.75699 (* 0.3 = 0.527096 loss) | |
| I0408 11:48:29.000280 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.466681 (* 0.3 = 0.140004 loss) | |
| I0408 11:48:29.000293 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.727273 | |
| I0408 11:48:29.000306 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 11:48:29.000319 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.909091 | |
| I0408 11:48:29.000332 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.00627 (* 1 = 1.00627 loss) | |
| I0408 11:48:29.000347 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.279032 (* 1 = 0.279032 loss) | |
| I0408 11:48:29.000360 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 11:48:29.000371 3443 solver.cpp:245] Train net output #16: total_confidence = 0.196153 | |
| I0408 11:48:29.000386 3443 sgd_solver.cpp:106] Iteration 63500, lr = 0.00909286 | |
| I0408 11:54:02.390983 3443 solver.cpp:229] Iteration 64000, loss = 3.42656 | |
| I0408 11:54:02.391273 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.137255 | |
| I0408 11:54:02.391291 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 11:54:02.391305 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.352941 | |
| I0408 11:54:02.391337 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.86649 (* 0.3 = 0.859946 loss) | |
| I0408 11:54:02.391355 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.86971 (* 0.3 = 0.260913 loss) | |
| I0408 11:54:02.391368 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.215686 | |
| I0408 11:54:02.391381 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.767045 | |
| I0408 11:54:02.391392 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.490196 | |
| I0408 11:54:02.391407 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.68894 (* 0.3 = 0.806681 loss) | |
| I0408 11:54:02.391422 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.80452 (* 0.3 = 0.241356 loss) | |
| I0408 11:54:02.391434 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.54902 | |
| I0408 11:54:02.391448 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 11:54:02.391463 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.803922 | |
| I0408 11:54:02.391477 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.72471 (* 1 = 1.72471 loss) | |
| I0408 11:54:02.391491 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.545151 (* 1 = 0.545151 loss) | |
| I0408 11:54:02.391505 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 11:54:02.391516 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0221892 | |
| I0408 11:54:02.391531 3443 sgd_solver.cpp:106] Iteration 64000, lr = 0.00908571 | |
| I0408 11:59:35.765523 3443 solver.cpp:229] Iteration 64500, loss = 3.4451 | |
| I0408 11:59:35.765673 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.261905 | |
| I0408 11:59:35.765693 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 11:59:35.765707 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.547619 | |
| I0408 11:59:35.765723 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.33548 (* 0.3 = 0.700643 loss) | |
| I0408 11:59:35.765738 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.693395 (* 0.3 = 0.208018 loss) | |
| I0408 11:59:35.765751 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.404762 | |
| I0408 11:59:35.765764 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 11:59:35.765776 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 11:59:35.765790 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.89033 (* 0.3 = 0.5671 loss) | |
| I0408 11:59:35.765805 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.556063 (* 0.3 = 0.166819 loss) | |
| I0408 11:59:35.765817 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.714286 | |
| I0408 11:59:35.765830 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0408 11:59:35.765842 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.857143 | |
| I0408 11:59:35.765856 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.12459 (* 1 = 1.12459 loss) | |
| I0408 11:59:35.765872 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.334134 (* 1 = 0.334134 loss) | |
| I0408 11:59:35.765883 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 11:59:35.765895 3443 solver.cpp:245] Train net output #16: total_confidence = 0.207497 | |
| I0408 11:59:35.765909 3443 sgd_solver.cpp:106] Iteration 64500, lr = 0.00907857 | |
| I0408 12:05:08.766448 3443 solver.cpp:338] Iteration 65000, Testing net (#0) | |
| I0408 12:05:49.576123 3443 solver.cpp:393] Test loss: 3.40687 | |
| I0408 12:05:49.576242 3443 solver.cpp:406] Test net output #0: loss1/accuracy = 0.209199 | |
| I0408 12:05:49.576262 3443 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.804818 | |
| I0408 12:05:49.576277 3443 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.467768 | |
| I0408 12:05:49.576292 3443 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.86446 (* 0.3 = 0.859338 loss) | |
| I0408 12:05:49.576308 3443 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.730046 (* 0.3 = 0.219014 loss) | |
| I0408 12:05:49.576319 3443 solver.cpp:406] Test net output #5: loss2/accuracy = 0.447484 | |
| I0408 12:05:49.576333 3443 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.860048 | |
| I0408 12:05:49.576344 3443 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.7481 | |
| I0408 12:05:49.576359 3443 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.88966 (* 0.3 = 0.566899 loss) | |
| I0408 12:05:49.576372 3443 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.490716 (* 0.3 = 0.147215 loss) | |
| I0408 12:05:49.576385 3443 solver.cpp:406] Test net output #10: loss3/accuracy = 0.670039 | |
| I0408 12:05:49.576397 3443 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.904183 | |
| I0408 12:05:49.576408 3443 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.851985 | |
| I0408 12:05:49.576422 3443 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.25658 (* 1 = 1.25658 loss) | |
| I0408 12:05:49.576436 3443 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.35783 (* 1 = 0.35783 loss) | |
| I0408 12:05:49.576449 3443 solver.cpp:406] Test net output #15: total_accuracy = 0.169 | |
| I0408 12:05:49.576460 3443 solver.cpp:406] Test net output #16: total_confidence = 0.113329 | |
| I0408 12:05:49.948267 3443 solver.cpp:229] Iteration 65000, loss = 3.40538 | |
| I0408 12:05:49.948309 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.207547 | |
| I0408 12:05:49.948326 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 12:05:49.948339 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.45283 | |
| I0408 12:05:49.948354 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.51305 (* 0.3 = 0.753916 loss) | |
| I0408 12:05:49.948369 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.781157 (* 0.3 = 0.234347 loss) | |
| I0408 12:05:49.948382 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.433962 | |
| I0408 12:05:49.948395 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 12:05:49.948407 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.698113 | |
| I0408 12:05:49.948421 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.03744 (* 0.3 = 0.611231 loss) | |
| I0408 12:05:49.948436 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.642785 (* 0.3 = 0.192836 loss) | |
| I0408 12:05:49.948453 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.622642 | |
| I0408 12:05:49.948467 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.869318 | |
| I0408 12:05:49.948479 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.830189 | |
| I0408 12:05:49.948493 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.19817 (* 1 = 1.19817 loss) | |
| I0408 12:05:49.948508 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.38789 (* 1 = 0.38789 loss) | |
| I0408 12:05:49.948519 3443 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 12:05:49.948531 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0922243 | |
| I0408 12:05:49.948546 3443 sgd_solver.cpp:106] Iteration 65000, lr = 0.00907143 | |
| I0408 12:11:23.205088 3443 solver.cpp:229] Iteration 65500, loss = 3.35518 | |
| I0408 12:11:23.205204 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.177778 | |
| I0408 12:11:23.205224 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 12:11:23.205237 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.444444 | |
| I0408 12:11:23.205255 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.5205 (* 0.3 = 0.75615 loss) | |
| I0408 12:11:23.205271 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.697258 (* 0.3 = 0.209177 loss) | |
| I0408 12:11:23.205283 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.288889 | |
| I0408 12:11:23.205297 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0408 12:11:23.205309 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.555556 | |
| I0408 12:11:23.205323 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.43026 (* 0.3 = 0.729079 loss) | |
| I0408 12:11:23.205338 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.701138 (* 0.3 = 0.210341 loss) | |
| I0408 12:11:23.205350 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.555556 | |
| I0408 12:11:23.205363 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0408 12:11:23.205375 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.822222 | |
| I0408 12:11:23.205389 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.58255 (* 1 = 1.58255 loss) | |
| I0408 12:11:23.205404 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.424321 (* 1 = 0.424321 loss) | |
| I0408 12:11:23.205416 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 12:11:23.205432 3443 solver.cpp:245] Train net output #16: total_confidence = 0.112649 | |
| I0408 12:11:23.205447 3443 sgd_solver.cpp:106] Iteration 65500, lr = 0.00906429 | |
| I0408 12:16:56.631533 3443 solver.cpp:229] Iteration 66000, loss = 3.31988 | |
| I0408 12:16:56.631906 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.266667 | |
| I0408 12:16:56.631938 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 12:16:56.631963 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.422222 | |
| I0408 12:16:56.631992 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.66012 (* 0.3 = 0.798035 loss) | |
| I0408 12:16:56.632024 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.742604 (* 0.3 = 0.222781 loss) | |
| I0408 12:16:56.632051 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.355556 | |
| I0408 12:16:56.632076 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 12:16:56.632099 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 12:16:56.632127 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.27552 (* 0.3 = 0.682655 loss) | |
| I0408 12:16:56.632153 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.61516 (* 0.3 = 0.184548 loss) | |
| I0408 12:16:56.632175 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.666667 | |
| I0408 12:16:56.632197 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0408 12:16:56.632220 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.822222 | |
| I0408 12:16:56.632246 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.1339 (* 1 = 1.1339 loss) | |
| I0408 12:16:56.632272 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.31551 (* 1 = 0.31551 loss) | |
| I0408 12:16:56.632297 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 12:16:56.632318 3443 solver.cpp:245] Train net output #16: total_confidence = 0.124892 | |
| I0408 12:16:56.632342 3443 sgd_solver.cpp:106] Iteration 66000, lr = 0.00905714 | |
| I0408 12:22:29.956660 3443 solver.cpp:229] Iteration 66500, loss = 3.32183 | |
| I0408 12:22:29.956926 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.183673 | |
| I0408 12:22:29.956946 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 12:22:29.956959 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.469388 | |
| I0408 12:22:29.956976 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.62551 (* 0.3 = 0.787652 loss) | |
| I0408 12:22:29.956992 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.786873 (* 0.3 = 0.236062 loss) | |
| I0408 12:22:29.957006 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.326531 | |
| I0408 12:22:29.957018 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 12:22:29.957031 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.693878 | |
| I0408 12:22:29.957044 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.02139 (* 0.3 = 0.606417 loss) | |
| I0408 12:22:29.957059 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.609497 (* 0.3 = 0.182849 loss) | |
| I0408 12:22:29.957072 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.77551 | |
| I0408 12:22:29.957083 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0408 12:22:29.957096 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.938776 | |
| I0408 12:22:29.957110 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.78081 (* 1 = 0.78081 loss) | |
| I0408 12:22:29.957125 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.236746 (* 1 = 0.236746 loss) | |
| I0408 12:22:29.957137 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 12:22:29.957149 3443 solver.cpp:245] Train net output #16: total_confidence = 0.0644862 | |
| I0408 12:22:29.957164 3443 sgd_solver.cpp:106] Iteration 66500, lr = 0.00905 | |
| I0408 12:28:03.351414 3443 solver.cpp:229] Iteration 67000, loss = 3.31722 | |
| I0408 12:28:03.351555 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.255814 | |
| I0408 12:28:03.351577 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 12:28:03.351589 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.534884 | |
| I0408 12:28:03.351608 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.44842 (* 0.3 = 0.734525 loss) | |
| I0408 12:28:03.351622 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.733326 (* 0.3 = 0.219998 loss) | |
| I0408 12:28:03.351635 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.488372 | |
| I0408 12:28:03.351649 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0408 12:28:03.351660 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.767442 | |
| I0408 12:28:03.351675 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.74245 (* 0.3 = 0.522735 loss) | |
| I0408 12:28:03.351689 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.525752 (* 0.3 = 0.157726 loss) | |
| I0408 12:28:03.351701 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.837209 | |
| I0408 12:28:03.351713 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0408 12:28:03.351727 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0408 12:28:03.351740 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.651136 (* 1 = 0.651136 loss) | |
| I0408 12:28:03.351754 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.202429 (* 1 = 0.202429 loss) | |
| I0408 12:28:03.351768 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 12:28:03.351779 3443 solver.cpp:245] Train net output #16: total_confidence = 0.166429 | |
| I0408 12:28:03.351794 3443 sgd_solver.cpp:106] Iteration 67000, lr = 0.00904286 | |
| I0408 12:33:36.740633 3443 solver.cpp:229] Iteration 67500, loss = 3.33852 | |
| I0408 12:33:36.740978 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0408 12:33:36.741006 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 12:33:36.741029 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.520833 | |
| I0408 12:33:36.741057 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.35802 (* 0.3 = 0.707407 loss) | |
| I0408 12:33:36.741086 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.708073 (* 0.3 = 0.212422 loss) | |
| I0408 12:33:36.741109 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.354167 | |
| I0408 12:33:36.741132 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0408 12:33:36.741156 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 12:33:36.741183 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.98254 (* 0.3 = 0.594763 loss) | |
| I0408 12:33:36.741207 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.587025 (* 0.3 = 0.176108 loss) | |
| I0408 12:33:36.741231 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.770833 | |
| I0408 12:33:36.741255 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 12:33:36.741281 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.854167 | |
| I0408 12:33:36.741308 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.920429 (* 1 = 0.920429 loss) | |
| I0408 12:33:36.741336 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.302467 (* 1 = 0.302467 loss) | |
| I0408 12:33:36.741358 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 12:33:36.741381 3443 solver.cpp:245] Train net output #16: total_confidence = 0.157872 | |
| I0408 12:33:36.741406 3443 sgd_solver.cpp:106] Iteration 67500, lr = 0.00903571 | |
| I0408 12:39:10.133369 3443 solver.cpp:229] Iteration 68000, loss = 3.32713 | |
| I0408 12:39:10.133566 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.2 | |
| I0408 12:39:10.133586 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 12:39:10.133600 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.42 | |
| I0408 12:39:10.133618 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.65689 (* 0.3 = 0.797066 loss) | |
| I0408 12:39:10.133633 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.79288 (* 0.3 = 0.237864 loss) | |
| I0408 12:39:10.133646 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.3 | |
| I0408 12:39:10.133659 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0408 12:39:10.133671 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.58 | |
| I0408 12:39:10.133687 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.38521 (* 0.3 = 0.715562 loss) | |
| I0408 12:39:10.133702 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.708181 (* 0.3 = 0.212454 loss) | |
| I0408 12:39:10.133714 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.66 | |
| I0408 12:39:10.133728 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 12:39:10.133739 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.86 | |
| I0408 12:39:10.133754 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.11182 (* 1 = 1.11182 loss) | |
| I0408 12:39:10.133769 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.38027 (* 1 = 0.38027 loss) | |
| I0408 12:39:10.133781 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 12:39:10.133793 3443 solver.cpp:245] Train net output #16: total_confidence = 0.237414 | |
| I0408 12:39:10.133808 3443 sgd_solver.cpp:106] Iteration 68000, lr = 0.00902857 | |
| I0408 12:44:43.525818 3443 solver.cpp:229] Iteration 68500, loss = 3.30194 | |
| I0408 12:44:43.526129 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.229167 | |
| I0408 12:44:43.526149 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 12:44:43.526163 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.416667 | |
| I0408 12:44:43.526180 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.53872 (* 0.3 = 0.761615 loss) | |
| I0408 12:44:43.526196 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.711161 (* 0.3 = 0.213348 loss) | |
| I0408 12:44:43.526208 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.291667 | |
| I0408 12:44:43.526221 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 12:44:43.526233 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.604167 | |
| I0408 12:44:43.526248 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.33546 (* 0.3 = 0.700638 loss) | |
| I0408 12:44:43.526262 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.689288 (* 0.3 = 0.206786 loss) | |
| I0408 12:44:43.526275 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.583333 | |
| I0408 12:44:43.526288 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 12:44:43.526300 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.791667 | |
| I0408 12:44:43.526315 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.38849 (* 1 = 1.38849 loss) | |
| I0408 12:44:43.526330 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.389279 (* 1 = 0.389279 loss) | |
| I0408 12:44:43.526342 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 12:44:43.526355 3443 solver.cpp:245] Train net output #16: total_confidence = 0.125763 | |
| I0408 12:44:43.526371 3443 sgd_solver.cpp:106] Iteration 68500, lr = 0.00902143 | |
| I0408 12:50:16.915913 3443 solver.cpp:229] Iteration 69000, loss = 3.2793 | |
| I0408 12:50:16.916113 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.166667 | |
| I0408 12:50:16.916136 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 12:50:16.916149 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.416667 | |
| I0408 12:50:16.916167 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.63101 (* 0.3 = 0.789303 loss) | |
| I0408 12:50:16.916182 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.777323 (* 0.3 = 0.233197 loss) | |
| I0408 12:50:16.916194 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.291667 | |
| I0408 12:50:16.916208 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 12:50:16.916219 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.645833 | |
| I0408 12:50:16.916234 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.24916 (* 0.3 = 0.674747 loss) | |
| I0408 12:50:16.916249 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.649462 (* 0.3 = 0.194839 loss) | |
| I0408 12:50:16.916261 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.729167 | |
| I0408 12:50:16.916273 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 12:50:16.916286 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.916667 | |
| I0408 12:50:16.916301 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.942724 (* 1 = 0.942724 loss) | |
| I0408 12:50:16.916316 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.282369 (* 1 = 0.282369 loss) | |
| I0408 12:50:16.916328 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 12:50:16.916340 3443 solver.cpp:245] Train net output #16: total_confidence = 0.234627 | |
| I0408 12:50:16.916357 3443 sgd_solver.cpp:106] Iteration 69000, lr = 0.00901429 | |
| I0408 12:55:50.322963 3443 solver.cpp:229] Iteration 69500, loss = 3.2878 | |
| I0408 12:55:50.323287 3443 solver.cpp:245] Train net output #0: loss1/accuracy = 0.292683 | |
| I0408 12:55:50.323307 3443 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 12:55:50.323338 3443 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.512195 | |
| I0408 12:55:50.323356 3443 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.48236 (* 0.3 = 0.744707 loss) | |
| I0408 12:55:50.323372 3443 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.716489 (* 0.3 = 0.214947 loss) | |
| I0408 12:55:50.323385 3443 solver.cpp:245] Train net output #5: loss2/accuracy = 0.292683 | |
| I0408 12:55:50.323398 3443 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 12:55:50.323410 3443 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.609756 | |
| I0408 12:55:50.323426 3443 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.21864 (* 0.3 = 0.665593 loss) | |
| I0408 12:55:50.323441 3443 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.696968 (* 0.3 = 0.20909 loss) | |
| I0408 12:55:50.323453 3443 solver.cpp:245] Train net output #10: loss3/accuracy = 0.560976 | |
| I0408 12:55:50.323467 3443 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.852273 | |
| I0408 12:55:50.323478 3443 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.756098 | |
| I0408 12:55:50.323493 3443 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.58245 (* 1 = 1.58245 loss) | |
| I0408 12:55:50.323506 3443 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.515651 (* 1 = 0.515651 loss) | |
| I0408 12:55:50.323519 3443 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 12:55:50.323532 3443 solver.cpp:245] Train net output #16: total_confidence = 0.143788 | |
| I0408 12:55:50.323546 3443 sgd_solver.cpp:106] Iteration 69500, lr = 0.00900714 | |
| I0408 15:42:06.473958 8707 solver.cpp:280] Solving mixed_lstm | |
| I0408 15:42:06.473971 8707 solver.cpp:281] Learning Rate Policy: poly | |
| I0408 15:42:06.494449 8707 solver.cpp:338] Iteration 70000, Testing net (#0) | |
| I0408 15:42:50.254896 8707 solver.cpp:393] Test loss: 3.14009 | |
| I0408 15:42:50.255269 8707 solver.cpp:406] Test net output #0: loss1/accuracy = 0.295435 | |
| I0408 15:42:50.255290 8707 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.823273 | |
| I0408 15:42:50.255302 8707 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.575922 | |
| I0408 15:42:50.255318 8707 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.46766 (* 0.3 = 0.740297 loss) | |
| I0408 15:42:50.255332 8707 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.640659 (* 0.3 = 0.192198 loss) | |
| I0408 15:42:50.255345 8707 solver.cpp:406] Test net output #5: loss2/accuracy = 0.497499 | |
| I0408 15:42:50.255357 8707 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.866458 | |
| I0408 15:42:50.255368 8707 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.756083 | |
| I0408 15:42:50.255383 8707 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.76814 (* 0.3 = 0.530441 loss) | |
| I0408 15:42:50.255395 8707 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.472837 (* 0.3 = 0.141851 loss) | |
| I0408 15:42:50.255408 8707 solver.cpp:406] Test net output #10: loss3/accuracy = 0.697533 | |
| I0408 15:42:50.255419 8707 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.909456 | |
| I0408 15:42:50.255431 8707 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.856439 | |
| I0408 15:42:50.255445 8707 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.19293 (* 1 = 1.19293 loss) | |
| I0408 15:42:50.255458 8707 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.342372 (* 1 = 0.342372 loss) | |
| I0408 15:42:50.255470 8707 solver.cpp:406] Test net output #15: total_accuracy = 0.147 | |
| I0408 15:42:50.255481 8707 solver.cpp:406] Test net output #16: total_confidence = 0.160926 | |
| I0408 15:42:50.977965 8707 solver.cpp:229] Iteration 70000, loss = 2.57209 | |
| I0408 15:42:50.978024 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.363636 | |
| I0408 15:42:50.978044 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0408 15:42:50.978056 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.590909 | |
| I0408 15:42:50.978075 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.15054 (* 0.3 = 0.645163 loss) | |
| I0408 15:42:50.978090 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.571997 (* 0.3 = 0.171599 loss) | |
| I0408 15:42:50.978102 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.431818 | |
| I0408 15:42:50.978116 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0408 15:42:50.978127 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.772727 | |
| I0408 15:42:50.978140 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.73462 (* 0.3 = 0.520385 loss) | |
| I0408 15:42:50.978155 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.479666 (* 0.3 = 0.1439 loss) | |
| I0408 15:42:50.978168 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.704545 | |
| I0408 15:42:50.978180 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0408 15:42:50.978193 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.909091 | |
| I0408 15:42:50.978206 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.845733 (* 1 = 0.845733 loss) | |
| I0408 15:42:50.978220 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.245311 (* 1 = 0.245311 loss) | |
| I0408 15:42:50.978234 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 15:42:50.978245 8707 solver.cpp:245] Train net output #16: total_confidence = 0.164924 | |
| I0408 15:42:50.978271 8707 sgd_solver.cpp:106] Iteration 70000, lr = 0.009 | |
| I0408 15:48:33.850759 8707 solver.cpp:229] Iteration 70500, loss = 3.29909 | |
| I0408 15:48:33.850924 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.348837 | |
| I0408 15:48:33.850946 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 15:48:33.850960 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.627907 | |
| I0408 15:48:33.850975 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.46495 (* 0.3 = 0.739485 loss) | |
| I0408 15:48:33.850991 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.864671 (* 0.3 = 0.259401 loss) | |
| I0408 15:48:33.851002 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.372093 | |
| I0408 15:48:33.851014 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 15:48:33.851027 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.697674 | |
| I0408 15:48:33.851040 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.11225 (* 0.3 = 0.633674 loss) | |
| I0408 15:48:33.851054 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.890119 (* 0.3 = 0.267036 loss) | |
| I0408 15:48:33.851066 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.72093 | |
| I0408 15:48:33.851078 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0408 15:48:33.851090 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0408 15:48:33.851104 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.831518 (* 1 = 0.831518 loss) | |
| I0408 15:48:33.851117 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.558633 (* 1 = 0.558633 loss) | |
| I0408 15:48:33.851130 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 15:48:33.851141 8707 solver.cpp:245] Train net output #16: total_confidence = 0.269043 | |
| I0408 15:48:33.851156 8707 sgd_solver.cpp:106] Iteration 70500, lr = 0.00899286 | |
| I0408 15:54:14.145403 8707 solver.cpp:229] Iteration 71000, loss = 3.26091 | |
| I0408 15:54:14.145546 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.191489 | |
| I0408 15:54:14.145566 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 15:54:14.145581 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.361702 | |
| I0408 15:54:14.145596 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.03794 (* 0.3 = 0.911382 loss) | |
| I0408 15:54:14.145612 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.874393 (* 0.3 = 0.262318 loss) | |
| I0408 15:54:14.145624 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.255319 | |
| I0408 15:54:14.145637 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0408 15:54:14.145648 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.489362 | |
| I0408 15:54:14.145663 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.88942 (* 0.3 = 0.866826 loss) | |
| I0408 15:54:14.145676 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.846134 (* 0.3 = 0.25384 loss) | |
| I0408 15:54:14.145689 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.617021 | |
| I0408 15:54:14.145701 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 15:54:14.145714 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.702128 | |
| I0408 15:54:14.145727 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.49237 (* 1 = 1.49237 loss) | |
| I0408 15:54:14.145741 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.469902 (* 1 = 0.469902 loss) | |
| I0408 15:54:14.145756 8707 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 15:54:14.145768 8707 solver.cpp:245] Train net output #16: total_confidence = 0.106006 | |
| I0408 15:54:14.145783 8707 sgd_solver.cpp:106] Iteration 71000, lr = 0.00898571 | |
| I0408 15:59:53.047387 8707 solver.cpp:229] Iteration 71500, loss = 3.25126 | |
| I0408 15:59:53.047528 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.361702 | |
| I0408 15:59:53.047547 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0408 15:59:53.047560 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.595745 | |
| I0408 15:59:53.047577 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.20013 (* 0.3 = 0.66004 loss) | |
| I0408 15:59:53.047591 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.65036 (* 0.3 = 0.195108 loss) | |
| I0408 15:59:53.047605 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.468085 | |
| I0408 15:59:53.047617 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0408 15:59:53.047629 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.787234 | |
| I0408 15:59:53.047643 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.71556 (* 0.3 = 0.514669 loss) | |
| I0408 15:59:53.047657 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.485297 (* 0.3 = 0.145589 loss) | |
| I0408 15:59:53.047669 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.765957 | |
| I0408 15:59:53.047682 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0408 15:59:53.047698 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0408 15:59:53.047725 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.741202 (* 1 = 0.741202 loss) | |
| I0408 15:59:53.047756 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.20162 (* 1 = 0.20162 loss) | |
| I0408 15:59:53.047780 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 15:59:53.047802 8707 solver.cpp:245] Train net output #16: total_confidence = 0.161017 | |
| I0408 15:59:53.047818 8707 sgd_solver.cpp:106] Iteration 71500, lr = 0.00897857 | |
| I0408 16:05:31.167095 8707 solver.cpp:229] Iteration 72000, loss = 3.24851 | |
| I0408 16:05:31.167230 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.288889 | |
| I0408 16:05:31.167250 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 16:05:31.167263 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.355556 | |
| I0408 16:05:31.167279 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.6579 (* 0.3 = 0.797369 loss) | |
| I0408 16:05:31.167294 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.786724 (* 0.3 = 0.236017 loss) | |
| I0408 16:05:31.167307 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4 | |
| I0408 16:05:31.167318 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0408 16:05:31.167330 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.533333 | |
| I0408 16:05:31.167345 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.28861 (* 0.3 = 0.686584 loss) | |
| I0408 16:05:31.167358 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.689335 (* 0.3 = 0.2068 loss) | |
| I0408 16:05:31.167371 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.488889 | |
| I0408 16:05:31.167383 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.852273 | |
| I0408 16:05:31.167395 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.733333 | |
| I0408 16:05:31.167409 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.02215 (* 1 = 2.02215 loss) | |
| I0408 16:05:31.167423 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.551221 (* 1 = 0.551221 loss) | |
| I0408 16:05:31.167436 8707 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 16:05:31.167448 8707 solver.cpp:245] Train net output #16: total_confidence = 0.0983394 | |
| I0408 16:05:31.167464 8707 sgd_solver.cpp:106] Iteration 72000, lr = 0.00897143 | |
| I0408 16:11:08.521605 8707 solver.cpp:229] Iteration 72500, loss = 3.22106 | |
| I0408 16:11:08.521821 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.243902 | |
| I0408 16:11:08.521842 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 16:11:08.521855 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.536585 | |
| I0408 16:11:08.521873 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.64527 (* 0.3 = 0.793581 loss) | |
| I0408 16:11:08.521888 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.683155 (* 0.3 = 0.204947 loss) | |
| I0408 16:11:08.521900 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.365854 | |
| I0408 16:11:08.521913 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0408 16:11:08.521924 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.658537 | |
| I0408 16:11:08.521939 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.09762 (* 0.3 = 0.629286 loss) | |
| I0408 16:11:08.521953 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.554902 (* 0.3 = 0.166471 loss) | |
| I0408 16:11:08.521965 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.609756 | |
| I0408 16:11:08.521977 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0408 16:11:08.521989 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.756098 | |
| I0408 16:11:08.522003 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.43715 (* 1 = 1.43715 loss) | |
| I0408 16:11:08.522017 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.371818 (* 1 = 0.371818 loss) | |
| I0408 16:11:08.522029 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 16:11:08.522042 8707 solver.cpp:245] Train net output #16: total_confidence = 0.161857 | |
| I0408 16:11:08.522056 8707 sgd_solver.cpp:106] Iteration 72500, lr = 0.00896429 | |
| I0408 16:16:45.525104 8707 solver.cpp:229] Iteration 73000, loss = 3.21028 | |
| I0408 16:16:45.525270 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.266667 | |
| I0408 16:16:45.525291 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 16:16:45.525305 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.555556 | |
| I0408 16:16:45.525321 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.48508 (* 0.3 = 0.745523 loss) | |
| I0408 16:16:45.525336 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.713707 (* 0.3 = 0.214112 loss) | |
| I0408 16:16:45.525349 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.422222 | |
| I0408 16:16:45.525362 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 16:16:45.525373 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.733333 | |
| I0408 16:16:45.525388 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.00921 (* 0.3 = 0.602762 loss) | |
| I0408 16:16:45.525401 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.591675 (* 0.3 = 0.177502 loss) | |
| I0408 16:16:45.525414 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.555556 | |
| I0408 16:16:45.525426 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0408 16:16:45.525439 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.822222 | |
| I0408 16:16:45.525454 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.60203 (* 1 = 1.60203 loss) | |
| I0408 16:16:45.525468 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.430654 (* 1 = 0.430654 loss) | |
| I0408 16:16:45.525480 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 16:16:45.525492 8707 solver.cpp:245] Train net output #16: total_confidence = 0.226538 | |
| I0408 16:16:45.525507 8707 sgd_solver.cpp:106] Iteration 73000, lr = 0.00895714 | |
| I0408 16:22:21.785189 8707 solver.cpp:229] Iteration 73500, loss = 3.20769 | |
| I0408 16:22:21.785377 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.184211 | |
| I0408 16:22:21.785398 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 16:22:21.785413 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.473684 | |
| I0408 16:22:21.785429 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.53334 (* 0.3 = 0.760001 loss) | |
| I0408 16:22:21.785444 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.613924 (* 0.3 = 0.184177 loss) | |
| I0408 16:22:21.785456 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.421053 | |
| I0408 16:22:21.785468 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0408 16:22:21.785480 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.763158 | |
| I0408 16:22:21.785495 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.91431 (* 0.3 = 0.574293 loss) | |
| I0408 16:22:21.785508 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.497319 (* 0.3 = 0.149196 loss) | |
| I0408 16:22:21.785521 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.763158 | |
| I0408 16:22:21.785533 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0408 16:22:21.785545 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.894737 | |
| I0408 16:22:21.785559 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.967538 (* 1 = 0.967538 loss) | |
| I0408 16:22:21.785573 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.220353 (* 1 = 0.220353 loss) | |
| I0408 16:22:21.785586 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 16:22:21.785598 8707 solver.cpp:245] Train net output #16: total_confidence = 0.253819 | |
| I0408 16:22:21.785614 8707 sgd_solver.cpp:106] Iteration 73500, lr = 0.00895 | |
| I0408 16:27:58.003021 8707 solver.cpp:229] Iteration 74000, loss = 3.16193 | |
| I0408 16:27:58.003190 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.230769 | |
| I0408 16:27:58.003212 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.755682 | |
| I0408 16:27:58.003226 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.442308 | |
| I0408 16:27:58.003242 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.6992 (* 0.3 = 0.809761 loss) | |
| I0408 16:27:58.003257 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.86977 (* 0.3 = 0.260931 loss) | |
| I0408 16:27:58.003269 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.403846 | |
| I0408 16:27:58.003283 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 16:27:58.003294 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.634615 | |
| I0408 16:27:58.003309 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.18173 (* 0.3 = 0.65452 loss) | |
| I0408 16:27:58.003324 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.751892 (* 0.3 = 0.225568 loss) | |
| I0408 16:27:58.003336 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.576923 | |
| I0408 16:27:58.003348 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.857955 | |
| I0408 16:27:58.003360 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.769231 | |
| I0408 16:27:58.003376 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.89972 (* 1 = 1.89972 loss) | |
| I0408 16:27:58.003389 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.628828 (* 1 = 0.628828 loss) | |
| I0408 16:27:58.003402 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 16:27:58.003414 8707 solver.cpp:245] Train net output #16: total_confidence = 0.206609 | |
| I0408 16:27:58.003430 8707 sgd_solver.cpp:106] Iteration 74000, lr = 0.00894286 | |
| I0408 16:33:33.534056 8707 solver.cpp:229] Iteration 74500, loss = 3.10243 | |
| I0408 16:33:33.534227 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.311111 | |
| I0408 16:33:33.534247 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 16:33:33.534260 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.533333 | |
| I0408 16:33:33.534276 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.13815 (* 0.3 = 0.641444 loss) | |
| I0408 16:33:33.534291 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.636352 (* 0.3 = 0.190905 loss) | |
| I0408 16:33:33.534304 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.6 | |
| I0408 16:33:33.534317 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0408 16:33:33.534328 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.866667 | |
| I0408 16:33:33.534343 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.34153 (* 0.3 = 0.402458 loss) | |
| I0408 16:33:33.534356 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.447047 (* 0.3 = 0.134114 loss) | |
| I0408 16:33:33.534369 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.688889 | |
| I0408 16:33:33.534381 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.903409 | |
| I0408 16:33:33.534394 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.888889 | |
| I0408 16:33:33.534406 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.933056 (* 1 = 0.933056 loss) | |
| I0408 16:33:33.534420 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.286905 (* 1 = 0.286905 loss) | |
| I0408 16:33:33.534432 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 16:33:33.534446 8707 solver.cpp:245] Train net output #16: total_confidence = 0.202984 | |
| I0408 16:33:33.534459 8707 sgd_solver.cpp:106] Iteration 74500, lr = 0.00893571 | |
| I0408 16:39:08.310617 8707 solver.cpp:338] Iteration 75000, Testing net (#0) | |
| I0408 16:39:49.685219 8707 solver.cpp:393] Test loss: 2.73925 | |
| I0408 16:39:49.685389 8707 solver.cpp:406] Test net output #0: loss1/accuracy = 0.360805 | |
| I0408 16:39:49.685410 8707 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.824683 | |
| I0408 16:39:49.685425 8707 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.665328 | |
| I0408 16:39:49.685439 8707 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.12212 (* 0.3 = 0.636636 loss) | |
| I0408 16:39:49.685454 8707 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.608654 (* 0.3 = 0.182596 loss) | |
| I0408 16:39:49.685467 8707 solver.cpp:406] Test net output #5: loss2/accuracy = 0.549683 | |
| I0408 16:39:49.685478 8707 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.852412 | |
| I0408 16:39:49.685489 8707 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.824926 | |
| I0408 16:39:49.685503 8707 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.55519 (* 0.3 = 0.466557 loss) | |
| I0408 16:39:49.685518 8707 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.50233 (* 0.3 = 0.150699 loss) | |
| I0408 16:39:49.685529 8707 solver.cpp:406] Test net output #10: loss3/accuracy = 0.738352 | |
| I0408 16:39:49.685540 8707 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.910774 | |
| I0408 16:39:49.685552 8707 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.89243 | |
| I0408 16:39:49.685565 8707 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.987608 (* 1 = 0.987608 loss) | |
| I0408 16:39:49.685580 8707 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.315155 (* 1 = 0.315155 loss) | |
| I0408 16:39:49.685590 8707 solver.cpp:406] Test net output #15: total_accuracy = 0.182 | |
| I0408 16:39:49.685609 8707 solver.cpp:406] Test net output #16: total_confidence = 0.170584 | |
| I0408 16:39:50.061269 8707 solver.cpp:229] Iteration 75000, loss = 3.16333 | |
| I0408 16:39:50.061326 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.208333 | |
| I0408 16:39:50.061343 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 16:39:50.061357 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.4375 | |
| I0408 16:39:50.061372 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.40978 (* 0.3 = 1.02294 loss) | |
| I0408 16:39:50.061388 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.00693 (* 0.3 = 0.302079 loss) | |
| I0408 16:39:50.061400 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.354167 | |
| I0408 16:39:50.061413 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 16:39:50.061424 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 16:39:50.061439 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.58006 (* 0.3 = 0.774018 loss) | |
| I0408 16:39:50.061452 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.770197 (* 0.3 = 0.231059 loss) | |
| I0408 16:39:50.061465 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.645833 | |
| I0408 16:39:50.061477 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 16:39:50.061488 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.770833 | |
| I0408 16:39:50.061502 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.5947 (* 1 = 2.5947 loss) | |
| I0408 16:39:50.061517 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.756216 (* 1 = 0.756216 loss) | |
| I0408 16:39:50.061529 8707 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 16:39:50.061542 8707 solver.cpp:245] Train net output #16: total_confidence = 0.190053 | |
| I0408 16:39:50.061558 8707 sgd_solver.cpp:106] Iteration 75000, lr = 0.00892857 | |
| I0408 16:45:24.709820 8707 solver.cpp:229] Iteration 75500, loss = 3.16236 | |
| I0408 16:45:24.710053 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.18 | |
| I0408 16:45:24.710072 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 16:45:24.710085 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.42 | |
| I0408 16:45:24.710101 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.78781 (* 0.3 = 0.836344 loss) | |
| I0408 16:45:24.710116 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.828207 (* 0.3 = 0.248462 loss) | |
| I0408 16:45:24.710129 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.32 | |
| I0408 16:45:24.710140 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 16:45:24.710152 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.58 | |
| I0408 16:45:24.710166 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.4962 (* 0.3 = 0.748859 loss) | |
| I0408 16:45:24.710181 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.731841 (* 0.3 = 0.219552 loss) | |
| I0408 16:45:24.710193 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.74 | |
| I0408 16:45:24.710206 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 16:45:24.710217 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9 | |
| I0408 16:45:24.710232 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.10157 (* 1 = 1.10157 loss) | |
| I0408 16:45:24.710245 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.336943 (* 1 = 0.336943 loss) | |
| I0408 16:45:24.710258 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 16:45:24.710269 8707 solver.cpp:245] Train net output #16: total_confidence = 0.161092 | |
| I0408 16:45:24.710286 8707 sgd_solver.cpp:106] Iteration 75500, lr = 0.00892143 | |
| I0408 16:50:59.409294 8707 solver.cpp:229] Iteration 76000, loss = 3.12183 | |
| I0408 16:50:59.409457 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.207547 | |
| I0408 16:50:59.409477 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0408 16:50:59.409490 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.509434 | |
| I0408 16:50:59.409507 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.44571 (* 0.3 = 0.733714 loss) | |
| I0408 16:50:59.409521 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.759755 (* 0.3 = 0.227926 loss) | |
| I0408 16:50:59.409534 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.45283 | |
| I0408 16:50:59.409546 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 16:50:59.409559 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.716981 | |
| I0408 16:50:59.409574 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.91889 (* 0.3 = 0.575667 loss) | |
| I0408 16:50:59.409589 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.608568 (* 0.3 = 0.18257 loss) | |
| I0408 16:50:59.409600 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.584906 | |
| I0408 16:50:59.409612 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.869318 | |
| I0408 16:50:59.409624 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.90566 | |
| I0408 16:50:59.409638 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.06587 (* 1 = 1.06587 loss) | |
| I0408 16:50:59.409652 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.339463 (* 1 = 0.339463 loss) | |
| I0408 16:50:59.409665 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 16:50:59.409677 8707 solver.cpp:245] Train net output #16: total_confidence = 0.0764779 | |
| I0408 16:50:59.409693 8707 sgd_solver.cpp:106] Iteration 76000, lr = 0.00891429 | |
| I0408 16:56:34.182046 8707 solver.cpp:229] Iteration 76500, loss = 3.11187 | |
| I0408 16:56:34.182325 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.2 | |
| I0408 16:56:34.182345 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0408 16:56:34.182359 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.422222 | |
| I0408 16:56:34.182376 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.61232 (* 0.3 = 0.783697 loss) | |
| I0408 16:56:34.182391 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.722049 (* 0.3 = 0.216615 loss) | |
| I0408 16:56:34.182404 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.422222 | |
| I0408 16:56:34.182416 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 16:56:34.182430 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.622222 | |
| I0408 16:56:34.182458 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.96692 (* 0.3 = 0.590077 loss) | |
| I0408 16:56:34.182479 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.552633 (* 0.3 = 0.16579 loss) | |
| I0408 16:56:34.182493 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.8 | |
| I0408 16:56:34.182505 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0408 16:56:34.182518 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.866667 | |
| I0408 16:56:34.182531 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.86918 (* 1 = 0.86918 loss) | |
| I0408 16:56:34.182545 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.249897 (* 1 = 0.249897 loss) | |
| I0408 16:56:34.182559 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 16:56:34.182570 8707 solver.cpp:245] Train net output #16: total_confidence = 0.155635 | |
| I0408 16:56:34.182585 8707 sgd_solver.cpp:106] Iteration 76500, lr = 0.00890714 | |
| I0408 17:02:08.899943 8707 solver.cpp:229] Iteration 77000, loss = 3.07348 | |
| I0408 17:02:08.900302 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.326087 | |
| I0408 17:02:08.900324 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 17:02:08.900338 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.565217 | |
| I0408 17:02:08.900355 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.44388 (* 0.3 = 0.733164 loss) | |
| I0408 17:02:08.900370 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.710086 (* 0.3 = 0.213026 loss) | |
| I0408 17:02:08.900383 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.369565 | |
| I0408 17:02:08.900395 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 17:02:08.900408 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.673913 | |
| I0408 17:02:08.900421 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.14927 (* 0.3 = 0.644781 loss) | |
| I0408 17:02:08.900435 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.631128 (* 0.3 = 0.189338 loss) | |
| I0408 17:02:08.900447 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.76087 | |
| I0408 17:02:08.900460 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0408 17:02:08.900473 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.891304 | |
| I0408 17:02:08.900511 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.00659 (* 1 = 1.00659 loss) | |
| I0408 17:02:08.900527 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.279815 (* 1 = 0.279815 loss) | |
| I0408 17:02:08.900538 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 17:02:08.900552 8707 solver.cpp:245] Train net output #16: total_confidence = 0.291625 | |
| I0408 17:02:08.900565 8707 sgd_solver.cpp:106] Iteration 77000, lr = 0.0089 | |
| I0408 17:07:43.256314 8707 solver.cpp:229] Iteration 77500, loss = 3.12204 | |
| I0408 17:07:43.256414 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.325581 | |
| I0408 17:07:43.256434 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 17:07:43.256448 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.581395 | |
| I0408 17:07:43.256464 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.31984 (* 0.3 = 0.695953 loss) | |
| I0408 17:07:43.256479 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.661303 (* 0.3 = 0.198391 loss) | |
| I0408 17:07:43.256506 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.511628 | |
| I0408 17:07:43.256520 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0408 17:07:43.256531 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.790698 | |
| I0408 17:07:43.256546 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52527 (* 0.3 = 0.457583 loss) | |
| I0408 17:07:43.256559 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.481942 (* 0.3 = 0.144583 loss) | |
| I0408 17:07:43.256572 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.837209 | |
| I0408 17:07:43.256584 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0408 17:07:43.256597 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.906977 | |
| I0408 17:07:43.256611 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.634815 (* 1 = 0.634815 loss) | |
| I0408 17:07:43.256625 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.184315 (* 1 = 0.184315 loss) | |
| I0408 17:07:43.256638 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 17:07:43.256649 8707 solver.cpp:245] Train net output #16: total_confidence = 0.246545 | |
| I0408 17:07:43.256664 8707 sgd_solver.cpp:106] Iteration 77500, lr = 0.00889286 | |
| I0408 17:13:17.327582 8707 solver.cpp:229] Iteration 78000, loss = 3.17777 | |
| I0408 17:13:17.327882 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.269231 | |
| I0408 17:13:17.327903 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 17:13:17.327915 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.480769 | |
| I0408 17:13:17.327932 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.54727 (* 0.3 = 0.764182 loss) | |
| I0408 17:13:17.327947 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.783228 (* 0.3 = 0.234968 loss) | |
| I0408 17:13:17.327960 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.384615 | |
| I0408 17:13:17.327973 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0408 17:13:17.327986 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.576923 | |
| I0408 17:13:17.327998 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.12143 (* 0.3 = 0.63643 loss) | |
| I0408 17:13:17.328013 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.64082 (* 0.3 = 0.192246 loss) | |
| I0408 17:13:17.328025 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.692308 | |
| I0408 17:13:17.328038 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.903409 | |
| I0408 17:13:17.328049 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.923077 | |
| I0408 17:13:17.328064 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.835142 (* 1 = 0.835142 loss) | |
| I0408 17:13:17.328078 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.269435 (* 1 = 0.269435 loss) | |
| I0408 17:13:17.328090 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 17:13:17.328104 8707 solver.cpp:245] Train net output #16: total_confidence = 0.104507 | |
| I0408 17:13:17.328117 8707 sgd_solver.cpp:106] Iteration 78000, lr = 0.00888571 | |
| I0408 17:18:52.079133 8707 solver.cpp:229] Iteration 78500, loss = 3.10331 | |
| I0408 17:18:52.079268 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.183673 | |
| I0408 17:18:52.079289 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0408 17:18:52.079305 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.469388 | |
| I0408 17:18:52.079322 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.49335 (* 0.3 = 0.748005 loss) | |
| I0408 17:18:52.079336 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.72026 (* 0.3 = 0.216078 loss) | |
| I0408 17:18:52.079349 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.387755 | |
| I0408 17:18:52.079362 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 17:18:52.079375 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.714286 | |
| I0408 17:18:52.079388 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.88002 (* 0.3 = 0.564005 loss) | |
| I0408 17:18:52.079402 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.549799 (* 0.3 = 0.16494 loss) | |
| I0408 17:18:52.079416 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.795918 | |
| I0408 17:18:52.079427 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0408 17:18:52.079439 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.918367 | |
| I0408 17:18:52.079453 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.905009 (* 1 = 0.905009 loss) | |
| I0408 17:18:52.079468 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.276912 (* 1 = 0.276912 loss) | |
| I0408 17:18:52.079480 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 17:18:52.079493 8707 solver.cpp:245] Train net output #16: total_confidence = 0.203936 | |
| I0408 17:18:52.079506 8707 sgd_solver.cpp:106] Iteration 78500, lr = 0.00887857 | |
| I0408 17:24:26.589975 8707 solver.cpp:229] Iteration 79000, loss = 3.06792 | |
| I0408 17:24:26.590353 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.380952 | |
| I0408 17:24:26.590374 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0408 17:24:26.590389 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.547619 | |
| I0408 17:24:26.590404 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.32372 (* 0.3 = 0.697117 loss) | |
| I0408 17:24:26.590420 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.635815 (* 0.3 = 0.190744 loss) | |
| I0408 17:24:26.590432 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.357143 | |
| I0408 17:24:26.590445 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 17:24:26.590456 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 17:24:26.590471 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.97177 (* 0.3 = 0.59153 loss) | |
| I0408 17:24:26.590486 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.515173 (* 0.3 = 0.154552 loss) | |
| I0408 17:24:26.590498 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.833333 | |
| I0408 17:24:26.590510 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0408 17:24:26.590523 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.952381 | |
| I0408 17:24:26.590536 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.632977 (* 1 = 0.632977 loss) | |
| I0408 17:24:26.590551 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.191864 (* 1 = 0.191864 loss) | |
| I0408 17:24:26.590564 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0408 17:24:26.590576 8707 solver.cpp:245] Train net output #16: total_confidence = 0.311084 | |
| I0408 17:24:26.590590 8707 sgd_solver.cpp:106] Iteration 79000, lr = 0.00887143 | |
| I0408 17:30:00.337054 8707 solver.cpp:229] Iteration 79500, loss = 3.07973 | |
| I0408 17:30:00.337147 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.319149 | |
| I0408 17:30:00.337165 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 17:30:00.337178 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.574468 | |
| I0408 17:30:00.337194 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.32078 (* 0.3 = 0.696233 loss) | |
| I0408 17:30:00.337210 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.663778 (* 0.3 = 0.199134 loss) | |
| I0408 17:30:00.337224 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.553191 | |
| I0408 17:30:00.337235 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0408 17:30:00.337249 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.829787 | |
| I0408 17:30:00.337265 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.46909 (* 0.3 = 0.440727 loss) | |
| I0408 17:30:00.337280 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.451567 (* 0.3 = 0.13547 loss) | |
| I0408 17:30:00.337292 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.87234 | |
| I0408 17:30:00.337304 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0408 17:30:00.337317 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.978723 | |
| I0408 17:30:00.337332 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.389131 (* 1 = 0.389131 loss) | |
| I0408 17:30:00.337345 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.117018 (* 1 = 0.117018 loss) | |
| I0408 17:30:00.337358 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0408 17:30:00.337370 8707 solver.cpp:245] Train net output #16: total_confidence = 0.304178 | |
| I0408 17:30:00.337384 8707 sgd_solver.cpp:106] Iteration 79500, lr = 0.00886429 | |
| I0408 17:35:33.257277 8707 solver.cpp:456] Snapshotting to binary proto file /mnt2/snapshots/1/mixed_lstm10_bn_iter_80000.caffemodel | |
| I0408 17:35:33.640089 8707 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt2/snapshots/1/mixed_lstm10_bn_iter_80000.solverstate | |
| I0408 17:35:33.828174 8707 solver.cpp:338] Iteration 80000, Testing net (#0) | |
| I0408 17:36:14.690225 8707 solver.cpp:393] Test loss: 2.99133 | |
| I0408 17:36:14.690338 8707 solver.cpp:406] Test net output #0: loss1/accuracy = 0.300511 | |
| I0408 17:36:14.690357 8707 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.822728 | |
| I0408 17:36:14.690371 8707 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.586804 | |
| I0408 17:36:14.690385 8707 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.36811 (* 0.3 = 0.710432 loss) | |
| I0408 17:36:14.690400 8707 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.613158 (* 0.3 = 0.183947 loss) | |
| I0408 17:36:14.690412 8707 solver.cpp:406] Test net output #5: loss2/accuracy = 0.547735 | |
| I0408 17:36:14.690424 8707 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.871866 | |
| I0408 17:36:14.690436 8707 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.812063 | |
| I0408 17:36:14.690450 8707 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.59676 (* 0.3 = 0.479028 loss) | |
| I0408 17:36:14.690464 8707 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.447835 (* 0.3 = 0.134351 loss) | |
| I0408 17:36:14.690475 8707 solver.cpp:406] Test net output #10: loss3/accuracy = 0.712425 | |
| I0408 17:36:14.690487 8707 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.925047 | |
| I0408 17:36:14.690498 8707 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.856386 | |
| I0408 17:36:14.690512 8707 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.17836 (* 1 = 1.17836 loss) | |
| I0408 17:36:14.690526 8707 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.305203 (* 1 = 0.305203 loss) | |
| I0408 17:36:14.690538 8707 solver.cpp:406] Test net output #15: total_accuracy = 0.333 | |
| I0408 17:36:14.690549 8707 solver.cpp:406] Test net output #16: total_confidence = 0.284053 | |
| I0408 17:36:15.063191 8707 solver.cpp:229] Iteration 80000, loss = 3.03264 | |
| I0408 17:36:15.063249 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.369565 | |
| I0408 17:36:15.063266 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 17:36:15.063279 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.695652 | |
| I0408 17:36:15.063297 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95854 (* 0.3 = 0.587561 loss) | |
| I0408 17:36:15.063311 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.592121 (* 0.3 = 0.177636 loss) | |
| I0408 17:36:15.063324 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.608696 | |
| I0408 17:36:15.063336 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0408 17:36:15.063350 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.804348 | |
| I0408 17:36:15.063364 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52657 (* 0.3 = 0.457972 loss) | |
| I0408 17:36:15.063380 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.463719 (* 0.3 = 0.139116 loss) | |
| I0408 17:36:15.063391 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.804348 | |
| I0408 17:36:15.063405 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0408 17:36:15.063416 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.913043 | |
| I0408 17:36:15.063431 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.63669 (* 1 = 0.63669 loss) | |
| I0408 17:36:15.063444 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.268516 (* 1 = 0.268516 loss) | |
| I0408 17:36:15.063457 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 17:36:15.063469 8707 solver.cpp:245] Train net output #16: total_confidence = 0.34347 | |
| I0408 17:36:15.063484 8707 sgd_solver.cpp:106] Iteration 80000, lr = 0.00885714 | |
| I0408 17:41:48.447929 8707 solver.cpp:229] Iteration 80500, loss = 3.05907 | |
| I0408 17:41:48.448273 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.22 | |
| I0408 17:41:48.448297 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 17:41:48.448309 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.38 | |
| I0408 17:41:48.448325 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.79009 (* 0.3 = 0.837028 loss) | |
| I0408 17:41:48.448340 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.849467 (* 0.3 = 0.25484 loss) | |
| I0408 17:41:48.448354 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.26 | |
| I0408 17:41:48.448365 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.784091 | |
| I0408 17:41:48.448377 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.5 | |
| I0408 17:41:48.448392 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.41334 (* 0.3 = 0.724001 loss) | |
| I0408 17:41:48.448407 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.717153 (* 0.3 = 0.215146 loss) | |
| I0408 17:41:48.448420 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.68 | |
| I0408 17:41:48.448431 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0408 17:41:48.448443 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.8 | |
| I0408 17:41:48.448457 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.33771 (* 1 = 1.33771 loss) | |
| I0408 17:41:48.448472 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.405535 (* 1 = 0.405535 loss) | |
| I0408 17:41:48.448501 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 17:41:48.448516 8707 solver.cpp:245] Train net output #16: total_confidence = 0.0934694 | |
| I0408 17:41:48.448531 8707 sgd_solver.cpp:106] Iteration 80500, lr = 0.00885 | |
| I0408 17:47:22.067620 8707 solver.cpp:229] Iteration 81000, loss = 3.04221 | |
| I0408 17:47:22.067775 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.22449 | |
| I0408 17:47:22.067796 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 17:47:22.067811 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.55102 | |
| I0408 17:47:22.067826 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.17574 (* 0.3 = 0.652722 loss) | |
| I0408 17:47:22.067842 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.634328 (* 0.3 = 0.190298 loss) | |
| I0408 17:47:22.067854 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.387755 | |
| I0408 17:47:22.067867 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 17:47:22.067878 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.795918 | |
| I0408 17:47:22.067893 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.72889 (* 0.3 = 0.518667 loss) | |
| I0408 17:47:22.067908 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.51399 (* 0.3 = 0.154197 loss) | |
| I0408 17:47:22.067919 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.877551 | |
| I0408 17:47:22.067931 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0408 17:47:22.067944 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.918367 | |
| I0408 17:47:22.067957 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.616591 (* 1 = 0.616591 loss) | |
| I0408 17:47:22.067971 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.182701 (* 1 = 0.182701 loss) | |
| I0408 17:47:22.067983 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0408 17:47:22.067996 8707 solver.cpp:245] Train net output #16: total_confidence = 0.233288 | |
| I0408 17:47:22.068009 8707 sgd_solver.cpp:106] Iteration 81000, lr = 0.00884286 | |
| I0408 17:52:55.462575 8707 solver.cpp:229] Iteration 81500, loss = 3.07769 | |
| I0408 17:52:55.462803 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.341463 | |
| I0408 17:52:55.462823 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0408 17:52:55.462836 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.536585 | |
| I0408 17:52:55.462852 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.35193 (* 0.3 = 0.70558 loss) | |
| I0408 17:52:55.462867 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.625225 (* 0.3 = 0.187567 loss) | |
| I0408 17:52:55.462880 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.487805 | |
| I0408 17:52:55.462893 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0408 17:52:55.462904 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.658537 | |
| I0408 17:52:55.462918 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.98074 (* 0.3 = 0.594223 loss) | |
| I0408 17:52:55.462932 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.52869 (* 0.3 = 0.158607 loss) | |
| I0408 17:52:55.462945 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.634146 | |
| I0408 17:52:55.462957 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0408 17:52:55.462970 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.853659 | |
| I0408 17:52:55.462985 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.17625 (* 1 = 1.17625 loss) | |
| I0408 17:52:55.462998 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.35196 (* 1 = 0.35196 loss) | |
| I0408 17:52:55.463011 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 17:52:55.463023 8707 solver.cpp:245] Train net output #16: total_confidence = 0.123438 | |
| I0408 17:52:55.463038 8707 sgd_solver.cpp:106] Iteration 81500, lr = 0.00883571 | |
| I0408 17:58:28.800583 8707 solver.cpp:229] Iteration 82000, loss = 2.987 | |
| I0408 17:58:28.800719 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.23913 | |
| I0408 17:58:28.800739 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 17:58:28.800755 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.478261 | |
| I0408 17:58:28.800771 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.53793 (* 0.3 = 0.761378 loss) | |
| I0408 17:58:28.800786 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.740443 (* 0.3 = 0.222133 loss) | |
| I0408 17:58:28.800799 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.347826 | |
| I0408 17:58:28.800812 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0408 17:58:28.800824 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.586957 | |
| I0408 17:58:28.800839 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.31162 (* 0.3 = 0.693486 loss) | |
| I0408 17:58:28.800853 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.673928 (* 0.3 = 0.202178 loss) | |
| I0408 17:58:28.800865 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.565217 | |
| I0408 17:58:28.800879 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.846591 | |
| I0408 17:58:28.800889 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.847826 | |
| I0408 17:58:28.800904 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.47288 (* 1 = 1.47288 loss) | |
| I0408 17:58:28.800918 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.539795 (* 1 = 0.539795 loss) | |
| I0408 17:58:28.800930 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0408 17:58:28.800942 8707 solver.cpp:245] Train net output #16: total_confidence = 0.0761612 | |
| I0408 17:58:28.800957 8707 sgd_solver.cpp:106] Iteration 82000, lr = 0.00882857 | |
| I0408 18:04:02.173528 8707 solver.cpp:229] Iteration 82500, loss = 3.09506 | |
| I0408 18:04:02.173770 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.326087 | |
| I0408 18:04:02.173790 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0408 18:04:02.173804 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.565217 | |
| I0408 18:04:02.173820 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.3998 (* 0.3 = 0.71994 loss) | |
| I0408 18:04:02.173835 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.684994 (* 0.3 = 0.205498 loss) | |
| I0408 18:04:02.173847 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.434783 | |
| I0408 18:04:02.173861 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0408 18:04:02.173871 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.73913 | |
| I0408 18:04:02.173885 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.76416 (* 0.3 = 0.529247 loss) | |
| I0408 18:04:02.173899 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.501512 (* 0.3 = 0.150454 loss) | |
| I0408 18:04:02.173913 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.695652 | |
| I0408 18:04:02.173924 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 18:04:02.173936 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.891304 | |
| I0408 18:04:02.173949 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.01186 (* 1 = 1.01186 loss) | |
| I0408 18:04:02.173964 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.282392 (* 1 = 0.282392 loss) | |
| I0408 18:04:02.173975 8707 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 18:04:02.173987 8707 solver.cpp:245] Train net output #16: total_confidence = 0.0786592 | |
| I0408 18:04:02.174001 8707 sgd_solver.cpp:106] Iteration 82500, lr = 0.00882143 | |
| I0408 18:09:35.552259 8707 solver.cpp:229] Iteration 83000, loss = 2.96551 | |
| I0408 18:09:35.552412 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.325 | |
| I0408 18:09:35.552433 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 18:09:35.552446 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.65 | |
| I0408 18:09:35.552462 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.22126 (* 0.3 = 0.666379 loss) | |
| I0408 18:09:35.552477 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.642284 (* 0.3 = 0.192685 loss) | |
| I0408 18:09:35.552490 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55 | |
| I0408 18:09:35.552503 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0408 18:09:35.552515 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.675 | |
| I0408 18:09:35.552530 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.9125 (* 0.3 = 0.573749 loss) | |
| I0408 18:09:35.552557 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.516975 (* 0.3 = 0.155093 loss) | |
| I0408 18:09:35.552570 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.85 | |
| I0408 18:09:35.552582 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0408 18:09:35.552594 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9 | |
| I0408 18:09:35.552608 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.832987 (* 1 = 0.832987 loss) | |
| I0408 18:09:35.552623 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.260665 (* 1 = 0.260665 loss) | |
| I0408 18:09:35.552634 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 18:09:35.552647 8707 solver.cpp:245] Train net output #16: total_confidence = 0.244681 | |
| I0408 18:09:35.552661 8707 sgd_solver.cpp:106] Iteration 83000, lr = 0.00881429 | |
| I0408 18:15:08.908161 8707 solver.cpp:229] Iteration 83500, loss = 2.97429 | |
| I0408 18:15:08.908427 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.215686 | |
| I0408 18:15:08.908448 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.744318 | |
| I0408 18:15:08.908462 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.490196 | |
| I0408 18:15:08.908478 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.83623 (* 0.3 = 0.850869 loss) | |
| I0408 18:15:08.908510 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.916341 (* 0.3 = 0.274902 loss) | |
| I0408 18:15:08.908524 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.411765 | |
| I0408 18:15:08.908537 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 18:15:08.908550 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 18:15:08.908563 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.06979 (* 0.3 = 0.620936 loss) | |
| I0408 18:15:08.908578 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.687369 (* 0.3 = 0.206211 loss) | |
| I0408 18:15:08.908591 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.686275 | |
| I0408 18:15:08.908604 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 18:15:08.908617 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.882353 | |
| I0408 18:15:08.908630 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.06502 (* 1 = 1.06502 loss) | |
| I0408 18:15:08.908644 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.355221 (* 1 = 0.355221 loss) | |
| I0408 18:15:08.908656 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 18:15:08.908669 8707 solver.cpp:245] Train net output #16: total_confidence = 0.133338 | |
| I0408 18:15:08.908684 8707 sgd_solver.cpp:106] Iteration 83500, lr = 0.00880714 | |
| I0408 18:20:42.273823 8707 solver.cpp:229] Iteration 84000, loss = 3.0311 | |
| I0408 18:20:42.273965 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.352941 | |
| I0408 18:20:42.273985 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0408 18:20:42.273998 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.529412 | |
| I0408 18:20:42.274014 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.20627 (* 0.3 = 0.661881 loss) | |
| I0408 18:20:42.274030 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.714417 (* 0.3 = 0.214325 loss) | |
| I0408 18:20:42.274042 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.470588 | |
| I0408 18:20:42.274055 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 18:20:42.274067 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72549 | |
| I0408 18:20:42.274080 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.89722 (* 0.3 = 0.569167 loss) | |
| I0408 18:20:42.274094 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.60295 (* 0.3 = 0.180885 loss) | |
| I0408 18:20:42.274106 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.901961 | |
| I0408 18:20:42.274119 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0408 18:20:42.274130 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.980392 | |
| I0408 18:20:42.274144 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.397409 (* 1 = 0.397409 loss) | |
| I0408 18:20:42.274158 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.122578 (* 1 = 0.122578 loss) | |
| I0408 18:20:42.274171 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0408 18:20:42.274183 8707 solver.cpp:245] Train net output #16: total_confidence = 0.260257 | |
| I0408 18:20:42.274197 8707 sgd_solver.cpp:106] Iteration 84000, lr = 0.0088 | |
| I0408 18:26:15.673602 8707 solver.cpp:229] Iteration 84500, loss = 3.0665 | |
| I0408 18:26:15.673862 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.219512 | |
| I0408 18:26:15.673884 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0408 18:26:15.673898 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.439024 | |
| I0408 18:26:15.673914 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.69846 (* 0.3 = 0.809538 loss) | |
| I0408 18:26:15.673930 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.786572 (* 0.3 = 0.235972 loss) | |
| I0408 18:26:15.673943 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.414634 | |
| I0408 18:26:15.673955 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0408 18:26:15.673967 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.536585 | |
| I0408 18:26:15.673981 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.25274 (* 0.3 = 0.675823 loss) | |
| I0408 18:26:15.673996 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.633057 (* 0.3 = 0.189917 loss) | |
| I0408 18:26:15.674008 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.658537 | |
| I0408 18:26:15.674021 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.903409 | |
| I0408 18:26:15.674033 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.804878 | |
| I0408 18:26:15.674047 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.06276 (* 1 = 1.06276 loss) | |
| I0408 18:26:15.674062 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.358029 (* 1 = 0.358029 loss) | |
| I0408 18:26:15.674073 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 18:26:15.674087 8707 solver.cpp:245] Train net output #16: total_confidence = 0.210731 | |
| I0408 18:26:15.674100 8707 sgd_solver.cpp:106] Iteration 84500, lr = 0.00879286 | |
| I0408 18:31:48.623544 8707 solver.cpp:338] Iteration 85000, Testing net (#0) | |
| I0408 18:32:29.498246 8707 solver.cpp:393] Test loss: 2.65124 | |
| I0408 18:32:29.498363 8707 solver.cpp:406] Test net output #0: loss1/accuracy = 0.339836 | |
| I0408 18:32:29.498381 8707 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.832728 | |
| I0408 18:32:29.498394 8707 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.644915 | |
| I0408 18:32:29.498410 8707 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.20487 (* 0.3 = 0.661462 loss) | |
| I0408 18:32:29.498425 8707 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.574947 (* 0.3 = 0.172484 loss) | |
| I0408 18:32:29.498437 8707 solver.cpp:406] Test net output #5: loss2/accuracy = 0.558076 | |
| I0408 18:32:29.498450 8707 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.881367 | |
| I0408 18:32:29.498461 8707 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.830814 | |
| I0408 18:32:29.498473 8707 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.51834 (* 0.3 = 0.455502 loss) | |
| I0408 18:32:29.498487 8707 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.414451 (* 0.3 = 0.124335 loss) | |
| I0408 18:32:29.498499 8707 solver.cpp:406] Test net output #10: loss3/accuracy = 0.751264 | |
| I0408 18:32:29.498512 8707 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.931229 | |
| I0408 18:32:29.498522 8707 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.888767 | |
| I0408 18:32:29.498536 8707 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.968915 (* 1 = 0.968915 loss) | |
| I0408 18:32:29.498549 8707 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.268544 (* 1 = 0.268544 loss) | |
| I0408 18:32:29.498561 8707 solver.cpp:406] Test net output #15: total_accuracy = 0.351 | |
| I0408 18:32:29.498574 8707 solver.cpp:406] Test net output #16: total_confidence = 0.254078 | |
| I0408 18:32:29.871906 8707 solver.cpp:229] Iteration 85000, loss = 2.98518 | |
| I0408 18:32:29.871970 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.386364 | |
| I0408 18:32:29.871989 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0408 18:32:29.872001 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.613636 | |
| I0408 18:32:29.872019 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.1325 (* 0.3 = 0.639751 loss) | |
| I0408 18:32:29.872033 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.60783 (* 0.3 = 0.182349 loss) | |
| I0408 18:32:29.872046 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.363636 | |
| I0408 18:32:29.872058 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0408 18:32:29.872071 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.704545 | |
| I0408 18:32:29.872084 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.93673 (* 0.3 = 0.58102 loss) | |
| I0408 18:32:29.872099 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.518482 (* 0.3 = 0.155545 loss) | |
| I0408 18:32:29.872112 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.704545 | |
| I0408 18:32:29.872123 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0408 18:32:29.872135 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.931818 | |
| I0408 18:32:29.872149 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.841591 (* 1 = 0.841591 loss) | |
| I0408 18:32:29.872164 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.232296 (* 1 = 0.232296 loss) | |
| I0408 18:32:29.872176 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 18:32:29.872189 8707 solver.cpp:245] Train net output #16: total_confidence = 0.110502 | |
| I0408 18:32:29.872202 8707 sgd_solver.cpp:106] Iteration 85000, lr = 0.00878571 | |
| I0408 18:38:03.208255 8707 solver.cpp:229] Iteration 85500, loss = 3.00995 | |
| I0408 18:38:03.208428 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.302326 | |
| I0408 18:38:03.208449 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0408 18:38:03.208462 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.697674 | |
| I0408 18:38:03.208478 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.19347 (* 0.3 = 0.658042 loss) | |
| I0408 18:38:03.208493 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.63722 (* 0.3 = 0.191166 loss) | |
| I0408 18:38:03.208506 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.372093 | |
| I0408 18:38:03.208518 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0408 18:38:03.208530 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.697674 | |
| I0408 18:38:03.208555 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.73769 (* 0.3 = 0.521308 loss) | |
| I0408 18:38:03.208573 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.54122 (* 0.3 = 0.162366 loss) | |
| I0408 18:38:03.208586 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.790698 | |
| I0408 18:38:03.208600 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0408 18:38:03.208611 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0408 18:38:03.208626 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.774108 (* 1 = 0.774108 loss) | |
| I0408 18:38:03.208639 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.248967 (* 1 = 0.248967 loss) | |
| I0408 18:38:03.208652 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0408 18:38:03.208663 8707 solver.cpp:245] Train net output #16: total_confidence = 0.235052 | |
| I0408 18:38:03.208678 8707 sgd_solver.cpp:106] Iteration 85500, lr = 0.00877857 | |
| I0408 18:43:36.439546 8707 solver.cpp:229] Iteration 86000, loss = 3.01417 | |
| I0408 18:43:36.439791 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.14 | |
| I0408 18:43:36.439810 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0408 18:43:36.439824 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.5 | |
| I0408 18:43:36.439841 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.59996 (* 0.3 = 0.779989 loss) | |
| I0408 18:43:36.439856 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.793021 (* 0.3 = 0.237906 loss) | |
| I0408 18:43:36.439868 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.36 | |
| I0408 18:43:36.439882 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0408 18:43:36.439893 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.58 | |
| I0408 18:43:36.439908 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.18678 (* 0.3 = 0.656035 loss) | |
| I0408 18:43:36.439921 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.665074 (* 0.3 = 0.199522 loss) | |
| I0408 18:43:36.439934 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.54 | |
| I0408 18:43:36.439946 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.846591 | |
| I0408 18:43:36.439957 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.84 | |
| I0408 18:43:36.439971 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.48096 (* 1 = 1.48096 loss) | |
| I0408 18:43:36.439985 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.4756 (* 1 = 0.4756 loss) | |
| I0408 18:43:36.439996 8707 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0408 18:43:36.440008 8707 solver.cpp:245] Train net output #16: total_confidence = 0.0568312 | |
| I0408 18:43:36.440022 8707 sgd_solver.cpp:106] Iteration 86000, lr = 0.00877143 | |
| I0408 18:49:09.819535 8707 solver.cpp:229] Iteration 86500, loss = 3.00121 | |
| I0408 18:49:09.819694 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.170732 | |
| I0408 18:49:09.819715 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0408 18:49:09.819728 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.487805 | |
| I0408 18:49:09.819747 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.73288 (* 0.3 = 0.819863 loss) | |
| I0408 18:49:09.819763 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.766004 (* 0.3 = 0.229801 loss) | |
| I0408 18:49:09.819777 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.365854 | |
| I0408 18:49:09.819788 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0408 18:49:09.819800 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.780488 | |
| I0408 18:49:09.819814 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.09337 (* 0.3 = 0.628012 loss) | |
| I0408 18:49:09.819829 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.572326 (* 0.3 = 0.171698 loss) | |
| I0408 18:49:09.819841 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.658537 | |
| I0408 18:49:09.819854 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0408 18:49:09.819865 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.878049 | |
| I0408 18:49:09.819880 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.09981 (* 1 = 1.09981 loss) | |
| I0408 18:49:09.819893 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.313241 (* 1 = 0.313241 loss) | |
| I0408 18:49:09.819905 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 18:49:09.819917 8707 solver.cpp:245] Train net output #16: total_confidence = 0.08579 | |
| I0408 18:49:09.819931 8707 sgd_solver.cpp:106] Iteration 86500, lr = 0.00876429 | |
| I0408 18:54:43.169487 8707 solver.cpp:229] Iteration 87000, loss = 3.01525 | |
| I0408 18:54:43.169713 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.282609 | |
| I0408 18:54:43.169731 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 18:54:43.169744 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.630435 | |
| I0408 18:54:43.169760 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.14713 (* 0.3 = 0.644138 loss) | |
| I0408 18:54:43.169775 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.648328 (* 0.3 = 0.194499 loss) | |
| I0408 18:54:43.169787 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.543478 | |
| I0408 18:54:43.169800 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0408 18:54:43.169812 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.847826 | |
| I0408 18:54:43.169826 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.53195 (* 0.3 = 0.459584 loss) | |
| I0408 18:54:43.169841 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.437659 (* 0.3 = 0.131298 loss) | |
| I0408 18:54:43.169852 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.76087 | |
| I0408 18:54:43.169865 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0408 18:54:43.169877 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0408 18:54:43.169891 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.709409 (* 1 = 0.709409 loss) | |
| I0408 18:54:43.169905 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.233434 (* 1 = 0.233434 loss) | |
| I0408 18:54:43.169917 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 18:54:43.169929 8707 solver.cpp:245] Train net output #16: total_confidence = 0.181273 | |
| I0408 18:54:43.169944 8707 sgd_solver.cpp:106] Iteration 87000, lr = 0.00875714 | |
| I0408 19:00:16.547564 8707 solver.cpp:229] Iteration 87500, loss = 2.87674 | |
| I0408 19:00:16.547816 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.372093 | |
| I0408 19:00:16.547838 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0408 19:00:16.547852 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.627907 | |
| I0408 19:00:16.547868 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.06953 (* 0.3 = 0.620859 loss) | |
| I0408 19:00:16.547883 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.599962 (* 0.3 = 0.179989 loss) | |
| I0408 19:00:16.547896 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55814 | |
| I0408 19:00:16.547909 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0408 19:00:16.547920 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.813953 | |
| I0408 19:00:16.547935 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.51036 (* 0.3 = 0.453107 loss) | |
| I0408 19:00:16.547950 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.446202 (* 0.3 = 0.133861 loss) | |
| I0408 19:00:16.547962 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.860465 | |
| I0408 19:00:16.547974 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0408 19:00:16.547986 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0408 19:00:16.548002 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.5362 (* 1 = 0.5362 loss) | |
| I0408 19:00:16.548015 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.159062 (* 1 = 0.159062 loss) | |
| I0408 19:00:16.548027 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0408 19:00:16.548039 8707 solver.cpp:245] Train net output #16: total_confidence = 0.438583 | |
| I0408 19:00:16.548054 8707 sgd_solver.cpp:106] Iteration 87500, lr = 0.00875 | |
| I0408 19:05:49.908521 8707 solver.cpp:229] Iteration 88000, loss = 2.94024 | |
| I0408 19:05:49.908767 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.367347 | |
| I0408 19:05:49.908788 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0408 19:05:49.908802 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.591837 | |
| I0408 19:05:49.908818 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.2037 (* 0.3 = 0.661111 loss) | |
| I0408 19:05:49.908833 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.636232 (* 0.3 = 0.19087 loss) | |
| I0408 19:05:49.908844 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.469388 | |
| I0408 19:05:49.908857 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0408 19:05:49.908869 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.77551 | |
| I0408 19:05:49.908882 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.57318 (* 0.3 = 0.471955 loss) | |
| I0408 19:05:49.908896 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.455267 (* 0.3 = 0.13658 loss) | |
| I0408 19:05:49.908908 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.77551 | |
| I0408 19:05:49.908921 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0408 19:05:49.908932 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.938776 | |
| I0408 19:05:49.908946 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.884549 (* 1 = 0.884549 loss) | |
| I0408 19:05:49.908960 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.255119 (* 1 = 0.255119 loss) | |
| I0408 19:05:49.908972 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0408 19:05:49.908984 8707 solver.cpp:245] Train net output #16: total_confidence = 0.297564 | |
| I0408 19:05:49.909001 8707 sgd_solver.cpp:106] Iteration 88000, lr = 0.00874286 | |
| I0408 19:11:23.286762 8707 solver.cpp:229] Iteration 88500, loss = 2.92946 | |
| I0408 19:11:23.286890 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.386364 | |
| I0408 19:11:23.286911 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0408 19:11:23.286923 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.659091 | |
| I0408 19:11:23.286939 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.9376 (* 0.3 = 0.581279 loss) | |
| I0408 19:11:23.286954 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.583042 (* 0.3 = 0.174913 loss) | |
| I0408 19:11:23.286968 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.522727 | |
| I0408 19:11:23.286980 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0408 19:11:23.286993 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.909091 | |
| I0408 19:11:23.287006 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.41161 (* 0.3 = 0.423482 loss) | |
| I0408 19:11:23.287021 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.411256 (* 0.3 = 0.123377 loss) | |
| I0408 19:11:23.287034 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.863636 | |
| I0408 19:11:23.287045 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0408 19:11:23.287057 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.954545 | |
| I0408 19:11:23.287071 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.576127 (* 1 = 0.576127 loss) | |
| I0408 19:11:23.287086 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.167656 (* 1 = 0.167656 loss) | |
| I0408 19:11:23.287097 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0408 19:11:23.287109 8707 solver.cpp:245] Train net output #16: total_confidence = 0.367608 | |
| I0408 19:11:23.287123 8707 sgd_solver.cpp:106] Iteration 88500, lr = 0.00873571 | |
| I0408 19:16:56.693526 8707 solver.cpp:229] Iteration 89000, loss = 2.88463 | |
| I0408 19:16:56.693789 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.244444 | |
| I0408 19:16:56.693810 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0408 19:16:56.693824 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.444444 | |
| I0408 19:16:56.693840 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.50412 (* 0.3 = 0.751237 loss) | |
| I0408 19:16:56.693855 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.744779 (* 0.3 = 0.223434 loss) | |
| I0408 19:16:56.693866 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.355556 | |
| I0408 19:16:56.693882 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0408 19:16:56.693895 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0408 19:16:56.693909 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.09439 (* 0.3 = 0.628317 loss) | |
| I0408 19:16:56.693924 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.673824 (* 0.3 = 0.202147 loss) | |
| I0408 19:16:56.693936 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.644444 | |
| I0408 19:16:56.693948 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0408 19:16:56.693960 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.866667 | |
| I0408 19:16:56.693974 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.17082 (* 1 = 1.17082 loss) | |
| I0408 19:16:56.693989 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.385109 (* 1 = 0.385109 loss) | |
| I0408 19:16:56.694000 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0408 19:16:56.694012 8707 solver.cpp:245] Train net output #16: total_confidence = 0.166263 | |
| I0408 19:16:56.694026 8707 sgd_solver.cpp:106] Iteration 89000, lr = 0.00872857 | |
| I0408 19:22:30.054682 8707 solver.cpp:229] Iteration 89500, loss = 2.94994 | |
| I0408 19:22:30.055043 8707 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0408 19:22:30.055065 8707 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0408 19:22:30.055078 8707 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.583333 | |
| I0408 19:22:30.055095 8707 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.37605 (* 0.3 = 0.712814 loss) | |
| I0408 19:22:30.055110 8707 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.733296 (* 0.3 = 0.219989 loss) | |
| I0408 19:22:30.055124 8707 solver.cpp:245] Train net output #5: loss2/accuracy = 0.416667 | |
| I0408 19:22:30.055136 8707 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0408 19:22:30.055148 8707 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.75 | |
| I0408 19:22:30.055162 8707 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.66825 (* 0.3 = 0.500475 loss) | |
| I0408 19:22:30.055176 8707 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.503162 (* 0.3 = 0.150949 loss) | |
| I0408 19:22:30.055188 8707 solver.cpp:245] Train net output #10: loss3/accuracy = 0.854167 | |
| I0408 19:22:30.055200 8707 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0408 19:22:30.055212 8707 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9375 | |
| I0408 19:22:30.055227 8707 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.551832 (* 1 = 0.551832 loss) | |
| I0408 19:22:30.055240 8707 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.158025 (* 1 = 0.158025 loss) | |
| I0408 19:22:30.055253 8707 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0408 19:22:30.055265 8707 solver.cpp:245] Train net output #16: total_confidence = 0.244925 | |
| I0408 19:22:30.055279 8707 sgd_solver.cpp:106] Iteration 89500, lr = 0.00872143 | |
| I0409 00:55:42.503006 12249 solver.cpp:280] Solving mixed_lstm | |
| I0409 00:55:42.503020 12249 solver.cpp:281] Learning Rate Policy: poly | |
| I0409 00:55:42.523172 12249 solver.cpp:338] Iteration 90000, Testing net (#0) | |
| I0409 00:56:23.569026 12249 solver.cpp:393] Test loss: 2.57982 | |
| I0409 00:56:23.569416 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.356605 | |
| I0409 00:56:23.569437 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.832274 | |
| I0409 00:56:23.569450 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.680637 | |
| I0409 00:56:23.569468 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.05094 (* 0.3 = 0.615282 loss) | |
| I0409 00:56:23.569483 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.554579 (* 0.3 = 0.166374 loss) | |
| I0409 00:56:23.569494 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.520773 | |
| I0409 00:56:23.569507 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.876367 | |
| I0409 00:56:23.569519 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.817653 | |
| I0409 00:56:23.569532 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.60275 (* 0.3 = 0.480825 loss) | |
| I0409 00:56:23.569545 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.421428 (* 0.3 = 0.126428 loss) | |
| I0409 00:56:23.569557 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.773594 | |
| I0409 00:56:23.569569 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.936456 | |
| I0409 00:56:23.569581 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.898757 | |
| I0409 00:56:23.569596 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.933398 (* 1 = 0.933398 loss) | |
| I0409 00:56:23.569608 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.257516 (* 1 = 0.257516 loss) | |
| I0409 00:56:23.569620 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.407 | |
| I0409 00:56:23.569633 12249 solver.cpp:406] Test net output #16: total_confidence = 0.341069 | |
| I0409 00:56:24.274955 12249 solver.cpp:229] Iteration 90000, loss = 3.04573 | |
| I0409 00:56:24.275022 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.27907 | |
| I0409 00:56:24.275039 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 00:56:24.275053 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.418605 | |
| I0409 00:56:24.275071 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.71841 (* 0.3 = 0.815523 loss) | |
| I0409 00:56:24.275087 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.867524 (* 0.3 = 0.260257 loss) | |
| I0409 00:56:24.275100 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.511628 | |
| I0409 00:56:24.275113 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 00:56:24.275125 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72093 | |
| I0409 00:56:24.275141 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.78847 (* 0.3 = 0.53654 loss) | |
| I0409 00:56:24.275156 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.583104 (* 0.3 = 0.174931 loss) | |
| I0409 00:56:24.275168 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.767442 | |
| I0409 00:56:24.275180 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 00:56:24.275193 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.860465 | |
| I0409 00:56:24.275207 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.938318 (* 1 = 0.938318 loss) | |
| I0409 00:56:24.275223 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.320164 (* 1 = 0.320164 loss) | |
| I0409 00:56:24.275234 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 00:56:24.275248 12249 solver.cpp:245] Train net output #16: total_confidence = 0.0990351 | |
| I0409 00:56:24.275274 12249 sgd_solver.cpp:106] Iteration 90000, lr = 0.00871429 | |
| I0409 01:01:57.473618 12249 solver.cpp:229] Iteration 90500, loss = 2.87117 | |
| I0409 01:01:57.473943 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.355556 | |
| I0409 01:01:57.473966 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 01:01:57.473979 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.666667 | |
| I0409 01:01:57.473996 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.0876 (* 0.3 = 0.62628 loss) | |
| I0409 01:01:57.474012 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.604099 (* 0.3 = 0.18123 loss) | |
| I0409 01:01:57.474025 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.466667 | |
| I0409 01:01:57.474036 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 01:01:57.474048 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.733333 | |
| I0409 01:01:57.474063 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.71163 (* 0.3 = 0.51349 loss) | |
| I0409 01:01:57.474077 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.51686 (* 0.3 = 0.155058 loss) | |
| I0409 01:01:57.474092 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.666667 | |
| I0409 01:01:57.474103 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 01:01:57.474115 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.933333 | |
| I0409 01:01:57.474129 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.14588 (* 1 = 1.14588 loss) | |
| I0409 01:01:57.474143 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.342521 (* 1 = 0.342521 loss) | |
| I0409 01:01:57.474155 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 01:01:57.474169 12249 solver.cpp:245] Train net output #16: total_confidence = 0.214732 | |
| I0409 01:01:57.474182 12249 sgd_solver.cpp:106] Iteration 90500, lr = 0.00870714 | |
| I0409 01:07:30.839711 12249 solver.cpp:229] Iteration 91000, loss = 2.9023 | |
| I0409 01:07:30.839815 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.159091 | |
| I0409 01:07:30.839834 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0409 01:07:30.839848 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.454545 | |
| I0409 01:07:30.839864 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.96345 (* 0.3 = 0.889036 loss) | |
| I0409 01:07:30.839879 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.84929 (* 0.3 = 0.254787 loss) | |
| I0409 01:07:30.839892 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.363636 | |
| I0409 01:07:30.839905 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 01:07:30.839917 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.75 | |
| I0409 01:07:30.839932 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.02524 (* 0.3 = 0.607573 loss) | |
| I0409 01:07:30.839947 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.617482 (* 0.3 = 0.185245 loss) | |
| I0409 01:07:30.839959 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.795455 | |
| I0409 01:07:30.839972 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 01:07:30.839984 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.954545 | |
| I0409 01:07:30.839998 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.82514 (* 1 = 0.82514 loss) | |
| I0409 01:07:30.840013 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.226472 (* 1 = 0.226472 loss) | |
| I0409 01:07:30.840025 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 01:07:30.840037 12249 solver.cpp:245] Train net output #16: total_confidence = 0.088946 | |
| I0409 01:07:30.840052 12249 sgd_solver.cpp:106] Iteration 91000, lr = 0.0087 | |
| I0409 01:13:04.547045 12249 solver.cpp:229] Iteration 91500, loss = 2.86614 | |
| I0409 01:13:04.547387 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.28 | |
| I0409 01:13:04.547417 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 01:13:04.547441 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.46 | |
| I0409 01:13:04.547472 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.65006 (* 0.3 = 0.79502 loss) | |
| I0409 01:13:04.547499 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.805304 (* 0.3 = 0.241591 loss) | |
| I0409 01:13:04.547523 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.36 | |
| I0409 01:13:04.547546 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 01:13:04.547574 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.56 | |
| I0409 01:13:04.547601 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.21016 (* 0.3 = 0.663047 loss) | |
| I0409 01:13:04.547631 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.664749 (* 0.3 = 0.199425 loss) | |
| I0409 01:13:04.547653 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.74 | |
| I0409 01:13:04.547677 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 01:13:04.547698 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.92 | |
| I0409 01:13:04.547724 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.06999 (* 1 = 1.06999 loss) | |
| I0409 01:13:04.547755 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.346564 (* 1 = 0.346564 loss) | |
| I0409 01:13:04.547780 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 01:13:04.547802 12249 solver.cpp:245] Train net output #16: total_confidence = 0.141045 | |
| I0409 01:13:04.547828 12249 sgd_solver.cpp:106] Iteration 91500, lr = 0.00869286 | |
| I0409 01:18:38.257272 12249 solver.cpp:229] Iteration 92000, loss = 2.87813 | |
| I0409 01:18:38.257385 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.222222 | |
| I0409 01:18:38.257414 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0409 01:18:38.257438 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.444444 | |
| I0409 01:18:38.257468 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.58191 (* 0.3 = 0.774572 loss) | |
| I0409 01:18:38.257495 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.837518 (* 0.3 = 0.251255 loss) | |
| I0409 01:18:38.257519 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.351852 | |
| I0409 01:18:38.257544 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0409 01:18:38.257570 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.648148 | |
| I0409 01:18:38.257597 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.87983 (* 0.3 = 0.563948 loss) | |
| I0409 01:18:38.257624 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.619257 (* 0.3 = 0.185777 loss) | |
| I0409 01:18:38.257647 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.796296 | |
| I0409 01:18:38.257668 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 01:18:38.257690 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.907407 | |
| I0409 01:18:38.257717 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.789904 (* 1 = 0.789904 loss) | |
| I0409 01:18:38.257748 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.278951 (* 1 = 0.278951 loss) | |
| I0409 01:18:38.257772 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 01:18:38.257794 12249 solver.cpp:245] Train net output #16: total_confidence = 0.179466 | |
| I0409 01:18:38.257820 12249 sgd_solver.cpp:106] Iteration 92000, lr = 0.00868571 | |
| I0409 01:24:11.618089 12249 solver.cpp:229] Iteration 92500, loss = 2.92217 | |
| I0409 01:24:11.618398 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.263158 | |
| I0409 01:24:11.618420 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 01:24:11.618434 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.473684 | |
| I0409 01:24:11.618451 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.8122 (* 0.3 = 0.84366 loss) | |
| I0409 01:24:11.618468 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.762948 (* 0.3 = 0.228884 loss) | |
| I0409 01:24:11.618480 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.289474 | |
| I0409 01:24:11.618494 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0409 01:24:11.618505 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.684211 | |
| I0409 01:24:11.618520 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.42719 (* 0.3 = 0.728157 loss) | |
| I0409 01:24:11.618533 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.741499 (* 0.3 = 0.22245 loss) | |
| I0409 01:24:11.618546 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.578947 | |
| I0409 01:24:11.618558 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0409 01:24:11.618571 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.789474 | |
| I0409 01:24:11.618584 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.61056 (* 1 = 1.61056 loss) | |
| I0409 01:24:11.618598 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.429304 (* 1 = 0.429304 loss) | |
| I0409 01:24:11.618612 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 01:24:11.618623 12249 solver.cpp:245] Train net output #16: total_confidence = 0.164802 | |
| I0409 01:24:11.618638 12249 sgd_solver.cpp:106] Iteration 92500, lr = 0.00867857 | |
| I0409 01:29:44.988615 12249 solver.cpp:229] Iteration 93000, loss = 2.91308 | |
| I0409 01:29:44.988739 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.116279 | |
| I0409 01:29:44.988759 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 01:29:44.988773 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.465116 | |
| I0409 01:29:44.988790 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.8333 (* 0.3 = 0.849989 loss) | |
| I0409 01:29:44.988806 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.781681 (* 0.3 = 0.234504 loss) | |
| I0409 01:29:44.988818 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.348837 | |
| I0409 01:29:44.988831 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 01:29:44.988843 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.697674 | |
| I0409 01:29:44.988857 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.17943 (* 0.3 = 0.653829 loss) | |
| I0409 01:29:44.988872 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.627782 (* 0.3 = 0.188335 loss) | |
| I0409 01:29:44.988884 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.697674 | |
| I0409 01:29:44.988896 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 01:29:44.988909 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.883721 | |
| I0409 01:29:44.988922 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.2899 (* 1 = 1.2899 loss) | |
| I0409 01:29:44.988936 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.384855 (* 1 = 0.384855 loss) | |
| I0409 01:29:44.988950 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 01:29:44.988961 12249 solver.cpp:245] Train net output #16: total_confidence = 0.204981 | |
| I0409 01:29:44.988976 12249 sgd_solver.cpp:106] Iteration 93000, lr = 0.00867143 | |
| I0409 01:35:18.363997 12249 solver.cpp:229] Iteration 93500, loss = 2.84844 | |
| I0409 01:35:18.364295 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0409 01:35:18.364315 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 01:35:18.364329 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.538462 | |
| I0409 01:35:18.364346 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.51957 (* 0.3 = 0.755872 loss) | |
| I0409 01:35:18.364362 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.76354 (* 0.3 = 0.229062 loss) | |
| I0409 01:35:18.364374 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.346154 | |
| I0409 01:35:18.364387 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0409 01:35:18.364398 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.711538 | |
| I0409 01:35:18.364413 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.87828 (* 0.3 = 0.563484 loss) | |
| I0409 01:35:18.364426 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.57162 (* 0.3 = 0.171486 loss) | |
| I0409 01:35:18.364439 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.692308 | |
| I0409 01:35:18.364450 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 01:35:18.364462 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.961538 | |
| I0409 01:35:18.364477 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.887355 (* 1 = 0.887355 loss) | |
| I0409 01:35:18.364511 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.266395 (* 1 = 0.266395 loss) | |
| I0409 01:35:18.364524 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 01:35:18.364537 12249 solver.cpp:245] Train net output #16: total_confidence = 0.158292 | |
| I0409 01:35:18.364552 12249 sgd_solver.cpp:106] Iteration 93500, lr = 0.00866429 | |
| I0409 01:40:51.735038 12249 solver.cpp:229] Iteration 94000, loss = 2.86169 | |
| I0409 01:40:51.735290 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.404762 | |
| I0409 01:40:51.735309 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 01:40:51.735322 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.619048 | |
| I0409 01:40:51.735339 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.06141 (* 0.3 = 0.618422 loss) | |
| I0409 01:40:51.735354 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.596282 (* 0.3 = 0.178885 loss) | |
| I0409 01:40:51.735368 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 01:40:51.735380 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 01:40:51.735393 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.833333 | |
| I0409 01:40:51.735406 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.48754 (* 0.3 = 0.446262 loss) | |
| I0409 01:40:51.735420 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.447557 (* 0.3 = 0.134267 loss) | |
| I0409 01:40:51.735433 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.809524 | |
| I0409 01:40:51.735445 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 01:40:51.735457 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.928571 | |
| I0409 01:40:51.735471 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.725544 (* 1 = 0.725544 loss) | |
| I0409 01:40:51.735486 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.201954 (* 1 = 0.201954 loss) | |
| I0409 01:40:51.735498 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 01:40:51.735510 12249 solver.cpp:245] Train net output #16: total_confidence = 0.43628 | |
| I0409 01:40:51.735524 12249 sgd_solver.cpp:106] Iteration 94000, lr = 0.00865714 | |
| I0409 01:46:25.101899 12249 solver.cpp:229] Iteration 94500, loss = 2.84423 | |
| I0409 01:46:25.102059 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.160714 | |
| I0409 01:46:25.102080 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.732955 | |
| I0409 01:46:25.102093 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.410714 | |
| I0409 01:46:25.102109 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.97296 (* 0.3 = 0.891889 loss) | |
| I0409 01:46:25.102125 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.958986 (* 0.3 = 0.287696 loss) | |
| I0409 01:46:25.102138 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.196429 | |
| I0409 01:46:25.102150 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.738636 | |
| I0409 01:46:25.102162 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.482143 | |
| I0409 01:46:25.102175 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 3.11189 (* 0.3 = 0.933566 loss) | |
| I0409 01:46:25.102190 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 1.03257 (* 0.3 = 0.309772 loss) | |
| I0409 01:46:25.102202 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.5 | |
| I0409 01:46:25.102215 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.840909 | |
| I0409 01:46:25.102226 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.75 | |
| I0409 01:46:25.102241 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.95866 (* 1 = 1.95866 loss) | |
| I0409 01:46:25.102254 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.635301 (* 1 = 0.635301 loss) | |
| I0409 01:46:25.102267 12249 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0409 01:46:25.102278 12249 solver.cpp:245] Train net output #16: total_confidence = 0.0565807 | |
| I0409 01:46:25.102293 12249 sgd_solver.cpp:106] Iteration 94500, lr = 0.00865 | |
| I0409 01:51:58.075446 12249 solver.cpp:338] Iteration 95000, Testing net (#0) | |
| I0409 01:52:39.116176 12249 solver.cpp:393] Test loss: 2.62675 | |
| I0409 01:52:39.116328 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.362579 | |
| I0409 01:52:39.116348 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.835729 | |
| I0409 01:52:39.116361 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.665097 | |
| I0409 01:52:39.116379 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.10976 (* 0.3 = 0.632928 loss) | |
| I0409 01:52:39.116394 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.562934 (* 0.3 = 0.16888 loss) | |
| I0409 01:52:39.116405 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.569629 | |
| I0409 01:52:39.116418 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.886003 | |
| I0409 01:52:39.116430 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.837462 | |
| I0409 01:52:39.116443 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.47874 (* 0.3 = 0.443622 loss) | |
| I0409 01:52:39.116457 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.398241 (* 0.3 = 0.119472 loss) | |
| I0409 01:52:39.116469 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.758965 | |
| I0409 01:52:39.116493 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.938364 | |
| I0409 01:52:39.116508 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.887183 | |
| I0409 01:52:39.116523 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.00269 (* 1 = 1.00269 loss) | |
| I0409 01:52:39.116538 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.259159 (* 1 = 0.259159 loss) | |
| I0409 01:52:39.116549 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.389 | |
| I0409 01:52:39.116560 12249 solver.cpp:406] Test net output #16: total_confidence = 0.302611 | |
| I0409 01:52:39.489275 12249 solver.cpp:229] Iteration 95000, loss = 2.78452 | |
| I0409 01:52:39.489338 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 01:52:39.489356 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 01:52:39.489369 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.555556 | |
| I0409 01:52:39.489385 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.19088 (* 0.3 = 0.657263 loss) | |
| I0409 01:52:39.489400 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.618245 (* 0.3 = 0.185473 loss) | |
| I0409 01:52:39.489413 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.511111 | |
| I0409 01:52:39.489426 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 01:52:39.489437 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.866667 | |
| I0409 01:52:39.489451 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.56835 (* 0.3 = 0.470504 loss) | |
| I0409 01:52:39.489465 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.47248 (* 0.3 = 0.141744 loss) | |
| I0409 01:52:39.489478 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.866667 | |
| I0409 01:52:39.489490 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 01:52:39.489502 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.933333 | |
| I0409 01:52:39.489516 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.524652 (* 1 = 0.524652 loss) | |
| I0409 01:52:39.489531 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.172419 (* 1 = 0.172419 loss) | |
| I0409 01:52:39.489542 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 01:52:39.489555 12249 solver.cpp:245] Train net output #16: total_confidence = 0.294216 | |
| I0409 01:52:39.489569 12249 sgd_solver.cpp:106] Iteration 95000, lr = 0.00864286 | |
| I0409 01:58:12.845011 12249 solver.cpp:229] Iteration 95500, loss = 2.80802 | |
| I0409 01:58:12.845156 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.282609 | |
| I0409 01:58:12.845176 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 01:58:12.845190 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.543478 | |
| I0409 01:58:12.845206 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.64437 (* 0.3 = 0.79331 loss) | |
| I0409 01:58:12.845221 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.764577 (* 0.3 = 0.229373 loss) | |
| I0409 01:58:12.845234 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.413043 | |
| I0409 01:58:12.845247 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0409 01:58:12.845259 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.73913 | |
| I0409 01:58:12.845273 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.80202 (* 0.3 = 0.540607 loss) | |
| I0409 01:58:12.845288 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.593826 (* 0.3 = 0.178148 loss) | |
| I0409 01:58:12.845300 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.869565 | |
| I0409 01:58:12.845312 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 01:58:12.845324 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.978261 | |
| I0409 01:58:12.845338 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.504891 (* 1 = 0.504891 loss) | |
| I0409 01:58:12.845353 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.149382 (* 1 = 0.149382 loss) | |
| I0409 01:58:12.845366 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 01:58:12.845377 12249 solver.cpp:245] Train net output #16: total_confidence = 0.200284 | |
| I0409 01:58:12.845392 12249 sgd_solver.cpp:106] Iteration 95500, lr = 0.00863571 | |
| I0409 02:03:46.218155 12249 solver.cpp:229] Iteration 96000, loss = 2.88656 | |
| I0409 02:03:46.218467 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.409091 | |
| I0409 02:03:46.218488 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 02:03:46.218502 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.681818 | |
| I0409 02:03:46.218519 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.93477 (* 0.3 = 0.58043 loss) | |
| I0409 02:03:46.218534 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.526081 (* 0.3 = 0.157824 loss) | |
| I0409 02:03:46.218547 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.545455 | |
| I0409 02:03:46.218559 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 02:03:46.218572 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.75 | |
| I0409 02:03:46.218586 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.65316 (* 0.3 = 0.495947 loss) | |
| I0409 02:03:46.218601 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.489519 (* 0.3 = 0.146856 loss) | |
| I0409 02:03:46.218613 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.840909 | |
| I0409 02:03:46.218626 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 02:03:46.218637 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977273 | |
| I0409 02:03:46.218652 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.449103 (* 1 = 0.449103 loss) | |
| I0409 02:03:46.218665 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.134463 (* 1 = 0.134463 loss) | |
| I0409 02:03:46.218678 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 02:03:46.218690 12249 solver.cpp:245] Train net output #16: total_confidence = 0.297701 | |
| I0409 02:03:46.218704 12249 sgd_solver.cpp:106] Iteration 96000, lr = 0.00862857 | |
| I0409 02:09:19.620496 12249 solver.cpp:229] Iteration 96500, loss = 2.88829 | |
| I0409 02:09:19.620647 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.369565 | |
| I0409 02:09:19.620667 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 02:09:19.620682 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.652174 | |
| I0409 02:09:19.620697 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.10894 (* 0.3 = 0.632683 loss) | |
| I0409 02:09:19.620713 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.6244 (* 0.3 = 0.18732 loss) | |
| I0409 02:09:19.620726 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.565217 | |
| I0409 02:09:19.620738 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 02:09:19.620754 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.913043 | |
| I0409 02:09:19.620769 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.21616 (* 0.3 = 0.364849 loss) | |
| I0409 02:09:19.620784 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.390896 (* 0.3 = 0.117269 loss) | |
| I0409 02:09:19.620796 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826087 | |
| I0409 02:09:19.620808 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 02:09:19.620820 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.978261 | |
| I0409 02:09:19.620836 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.44003 (* 1 = 0.44003 loss) | |
| I0409 02:09:19.620849 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.147872 (* 1 = 0.147872 loss) | |
| I0409 02:09:19.620862 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 02:09:19.620873 12249 solver.cpp:245] Train net output #16: total_confidence = 0.33311 | |
| I0409 02:09:19.620889 12249 sgd_solver.cpp:106] Iteration 96500, lr = 0.00862143 | |
| I0409 02:14:52.968894 12249 solver.cpp:229] Iteration 97000, loss = 2.81622 | |
| I0409 02:14:52.969243 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.276596 | |
| I0409 02:14:52.969265 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 02:14:52.969279 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.489362 | |
| I0409 02:14:52.969296 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.59669 (* 0.3 = 0.779008 loss) | |
| I0409 02:14:52.969312 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.771925 (* 0.3 = 0.231578 loss) | |
| I0409 02:14:52.969324 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.446809 | |
| I0409 02:14:52.969337 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 02:14:52.969350 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.702128 | |
| I0409 02:14:52.969364 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.78672 (* 0.3 = 0.536017 loss) | |
| I0409 02:14:52.969378 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.569903 (* 0.3 = 0.170971 loss) | |
| I0409 02:14:52.969391 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.723404 | |
| I0409 02:14:52.969403 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 02:14:52.969415 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.851064 | |
| I0409 02:14:52.969430 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.929496 (* 1 = 0.929496 loss) | |
| I0409 02:14:52.969444 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.327139 (* 1 = 0.327139 loss) | |
| I0409 02:14:52.969457 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 02:14:52.969470 12249 solver.cpp:245] Train net output #16: total_confidence = 0.224099 | |
| I0409 02:14:52.969485 12249 sgd_solver.cpp:106] Iteration 97000, lr = 0.00861429 | |
| I0409 02:20:26.333663 12249 solver.cpp:229] Iteration 97500, loss = 2.81084 | |
| I0409 02:20:26.333961 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.339286 | |
| I0409 02:20:26.333982 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 02:20:26.333997 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.553571 | |
| I0409 02:20:26.334013 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.42013 (* 0.3 = 0.726038 loss) | |
| I0409 02:20:26.334028 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.815888 (* 0.3 = 0.244767 loss) | |
| I0409 02:20:26.334041 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.482143 | |
| I0409 02:20:26.334053 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 02:20:26.334066 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.696429 | |
| I0409 02:20:26.334080 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.83807 (* 0.3 = 0.551421 loss) | |
| I0409 02:20:26.334095 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.599152 (* 0.3 = 0.179746 loss) | |
| I0409 02:20:26.334107 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.75 | |
| I0409 02:20:26.334120 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 02:20:26.334131 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.875 | |
| I0409 02:20:26.334146 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.01354 (* 1 = 1.01354 loss) | |
| I0409 02:20:26.334161 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.329275 (* 1 = 0.329275 loss) | |
| I0409 02:20:26.334172 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 02:20:26.334184 12249 solver.cpp:245] Train net output #16: total_confidence = 0.346715 | |
| I0409 02:20:26.334199 12249 sgd_solver.cpp:106] Iteration 97500, lr = 0.00860714 | |
| I0409 02:26:00.044270 12249 solver.cpp:229] Iteration 98000, loss = 2.8159 | |
| I0409 02:26:00.044405 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.263158 | |
| I0409 02:26:00.044425 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0409 02:26:00.044440 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.45614 | |
| I0409 02:26:00.044456 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.66737 (* 0.3 = 0.80021 loss) | |
| I0409 02:26:00.044471 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.882365 (* 0.3 = 0.26471 loss) | |
| I0409 02:26:00.044483 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.315789 | |
| I0409 02:26:00.044495 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.778409 | |
| I0409 02:26:00.044507 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.701754 | |
| I0409 02:26:00.044522 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.16277 (* 0.3 = 0.648832 loss) | |
| I0409 02:26:00.044549 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.709709 (* 0.3 = 0.212913 loss) | |
| I0409 02:26:00.044564 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.736842 | |
| I0409 02:26:00.044575 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 02:26:00.044587 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.894737 | |
| I0409 02:26:00.044601 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.877061 (* 1 = 0.877061 loss) | |
| I0409 02:26:00.044615 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.293708 (* 1 = 0.293708 loss) | |
| I0409 02:26:00.044628 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 02:26:00.044641 12249 solver.cpp:245] Train net output #16: total_confidence = 0.14613 | |
| I0409 02:26:00.044654 12249 sgd_solver.cpp:106] Iteration 98000, lr = 0.0086 | |
| I0409 02:31:33.411733 12249 solver.cpp:229] Iteration 98500, loss = 2.82857 | |
| I0409 02:31:33.411934 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.416667 | |
| I0409 02:31:33.411955 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 02:31:33.411968 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.583333 | |
| I0409 02:31:33.411984 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.06952 (* 0.3 = 0.620857 loss) | |
| I0409 02:31:33.411999 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.628607 (* 0.3 = 0.188582 loss) | |
| I0409 02:31:33.412011 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 02:31:33.412024 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 02:31:33.412035 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.8125 | |
| I0409 02:31:33.412050 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52863 (* 0.3 = 0.458588 loss) | |
| I0409 02:31:33.412065 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.467488 (* 0.3 = 0.140247 loss) | |
| I0409 02:31:33.412077 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.854167 | |
| I0409 02:31:33.412089 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 02:31:33.412101 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.958333 | |
| I0409 02:31:33.412116 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.560507 (* 1 = 0.560507 loss) | |
| I0409 02:31:33.412129 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.162209 (* 1 = 0.162209 loss) | |
| I0409 02:31:33.412142 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 02:31:33.412153 12249 solver.cpp:245] Train net output #16: total_confidence = 0.366749 | |
| I0409 02:31:33.412168 12249 sgd_solver.cpp:106] Iteration 98500, lr = 0.00859286 | |
| I0409 02:37:06.783778 12249 solver.cpp:229] Iteration 99000, loss = 2.80202 | |
| I0409 02:37:06.783937 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 02:37:06.783958 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 02:37:06.783972 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.5 | |
| I0409 02:37:06.783989 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.57052 (* 0.3 = 0.771156 loss) | |
| I0409 02:37:06.784004 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.730033 (* 0.3 = 0.21901 loss) | |
| I0409 02:37:06.784016 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.325 | |
| I0409 02:37:06.784029 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0409 02:37:06.784041 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.625 | |
| I0409 02:37:06.784055 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.10448 (* 0.3 = 0.631344 loss) | |
| I0409 02:37:06.784070 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.687101 (* 0.3 = 0.20613 loss) | |
| I0409 02:37:06.784081 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.675 | |
| I0409 02:37:06.784095 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 02:37:06.784106 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.8 | |
| I0409 02:37:06.784121 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.26241 (* 1 = 1.26241 loss) | |
| I0409 02:37:06.784135 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.371476 (* 1 = 0.371476 loss) | |
| I0409 02:37:06.784147 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 02:37:06.784159 12249 solver.cpp:245] Train net output #16: total_confidence = 0.220528 | |
| I0409 02:37:06.784173 12249 sgd_solver.cpp:106] Iteration 99000, lr = 0.00858571 | |
| I0409 02:42:40.514129 12249 solver.cpp:229] Iteration 99500, loss = 2.77062 | |
| I0409 02:42:40.514415 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.395833 | |
| I0409 02:42:40.514446 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 02:42:40.514468 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.604167 | |
| I0409 02:42:40.514497 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.17829 (* 0.3 = 0.653487 loss) | |
| I0409 02:42:40.514523 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.694461 (* 0.3 = 0.208338 loss) | |
| I0409 02:42:40.514546 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.416667 | |
| I0409 02:42:40.514569 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 02:42:40.514590 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.645833 | |
| I0409 02:42:40.514614 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.93311 (* 0.3 = 0.579932 loss) | |
| I0409 02:42:40.514639 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.589271 (* 0.3 = 0.176781 loss) | |
| I0409 02:42:40.514662 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.75 | |
| I0409 02:42:40.514684 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 02:42:40.514708 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.833333 | |
| I0409 02:42:40.514735 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.05243 (* 1 = 1.05243 loss) | |
| I0409 02:42:40.514761 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.307638 (* 1 = 0.307638 loss) | |
| I0409 02:42:40.514783 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 02:42:40.514804 12249 solver.cpp:245] Train net output #16: total_confidence = 0.422808 | |
| I0409 02:42:40.514829 12249 sgd_solver.cpp:106] Iteration 99500, lr = 0.00857857 | |
| I0409 02:48:13.805173 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_100000.caffemodel | |
| I0409 02:48:14.247494 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_100000.solverstate | |
| I0409 02:48:14.487318 12249 solver.cpp:338] Iteration 100000, Testing net (#0) | |
| I0409 02:48:56.345772 12249 solver.cpp:393] Test loss: 2.47916 | |
| I0409 02:48:56.345862 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.389447 | |
| I0409 02:48:56.345880 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.833638 | |
| I0409 02:48:56.345893 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.699179 | |
| I0409 02:48:56.345909 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.01544 (* 0.3 = 0.604633 loss) | |
| I0409 02:48:56.345924 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.566369 (* 0.3 = 0.169911 loss) | |
| I0409 02:48:56.345937 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.598308 | |
| I0409 02:48:56.345948 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.887048 | |
| I0409 02:48:56.345960 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.851232 | |
| I0409 02:48:56.345974 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.40007 (* 0.3 = 0.420022 loss) | |
| I0409 02:48:56.345988 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.39165 (* 0.3 = 0.117495 loss) | |
| I0409 02:48:56.346000 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.780273 | |
| I0409 02:48:56.346012 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.929865 | |
| I0409 02:48:56.346024 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.899249 | |
| I0409 02:48:56.346037 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.894687 (* 1 = 0.894687 loss) | |
| I0409 02:48:56.346051 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.272417 (* 1 = 0.272417 loss) | |
| I0409 02:48:56.346062 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.368 | |
| I0409 02:48:56.346074 12249 solver.cpp:406] Test net output #16: total_confidence = 0.316005 | |
| I0409 02:48:56.719123 12249 solver.cpp:229] Iteration 100000, loss = 2.82872 | |
| I0409 02:48:56.719175 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 02:48:56.719193 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 02:48:56.719207 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.525 | |
| I0409 02:48:56.719223 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.31507 (* 0.3 = 0.694522 loss) | |
| I0409 02:48:56.719238 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.610244 (* 0.3 = 0.183073 loss) | |
| I0409 02:48:56.719251 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.425 | |
| I0409 02:48:56.719264 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 02:48:56.719280 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.725 | |
| I0409 02:48:56.719293 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.8754 (* 0.3 = 0.562621 loss) | |
| I0409 02:48:56.719307 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.495721 (* 0.3 = 0.148716 loss) | |
| I0409 02:48:56.719321 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.675 | |
| I0409 02:48:56.719332 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 02:48:56.719344 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.85 | |
| I0409 02:48:56.719359 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.05928 (* 1 = 1.05928 loss) | |
| I0409 02:48:56.719373 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.277421 (* 1 = 0.277421 loss) | |
| I0409 02:48:56.719385 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 02:48:56.719398 12249 solver.cpp:245] Train net output #16: total_confidence = 0.161994 | |
| I0409 02:48:56.719413 12249 sgd_solver.cpp:106] Iteration 100000, lr = 0.00857143 | |
| I0409 02:54:30.970578 12249 solver.cpp:229] Iteration 100500, loss = 2.75407 | |
| I0409 02:54:30.970973 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.326087 | |
| I0409 02:54:30.970998 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 02:54:30.971011 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.543478 | |
| I0409 02:54:30.971029 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.42057 (* 0.3 = 0.72617 loss) | |
| I0409 02:54:30.971051 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.699398 (* 0.3 = 0.20982 loss) | |
| I0409 02:54:30.971072 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.434783 | |
| I0409 02:54:30.971096 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 02:54:30.971119 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.76087 | |
| I0409 02:54:30.971138 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.76756 (* 0.3 = 0.530268 loss) | |
| I0409 02:54:30.971153 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.529264 (* 0.3 = 0.158779 loss) | |
| I0409 02:54:30.971165 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.847826 | |
| I0409 02:54:30.971177 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 02:54:30.971189 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 02:54:30.971204 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.84508 (* 1 = 0.84508 loss) | |
| I0409 02:54:30.971217 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.331098 (* 1 = 0.331098 loss) | |
| I0409 02:54:30.971230 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 02:54:30.971242 12249 solver.cpp:245] Train net output #16: total_confidence = 0.399372 | |
| I0409 02:54:30.971257 12249 sgd_solver.cpp:106] Iteration 100500, lr = 0.00856429 | |
| I0409 03:00:04.994613 12249 solver.cpp:229] Iteration 101000, loss = 2.8018 | |
| I0409 03:00:04.994839 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 03:00:04.994860 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 03:00:04.994884 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.45 | |
| I0409 03:00:04.994900 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.27406 (* 0.3 = 0.682219 loss) | |
| I0409 03:00:04.994915 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.682407 (* 0.3 = 0.204722 loss) | |
| I0409 03:00:04.994930 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.425 | |
| I0409 03:00:04.994941 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 03:00:04.994953 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.75 | |
| I0409 03:00:04.994967 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.72047 (* 0.3 = 0.516141 loss) | |
| I0409 03:00:04.994982 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.580313 (* 0.3 = 0.174094 loss) | |
| I0409 03:00:04.994994 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.7 | |
| I0409 03:00:04.995007 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 03:00:04.995019 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.875 | |
| I0409 03:00:04.995033 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.05484 (* 1 = 1.05484 loss) | |
| I0409 03:00:04.995048 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.351235 (* 1 = 0.351235 loss) | |
| I0409 03:00:04.995059 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 03:00:04.995071 12249 solver.cpp:245] Train net output #16: total_confidence = 0.324518 | |
| I0409 03:00:04.995087 12249 sgd_solver.cpp:106] Iteration 101000, lr = 0.00855714 | |
| I0409 03:05:39.993187 12249 solver.cpp:229] Iteration 101500, loss = 2.76511 | |
| I0409 03:05:39.993412 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.26 | |
| I0409 03:05:39.993430 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 03:05:39.993443 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.56 | |
| I0409 03:05:39.993461 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.23355 (* 0.3 = 0.670066 loss) | |
| I0409 03:05:39.993476 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.668353 (* 0.3 = 0.200506 loss) | |
| I0409 03:05:39.993489 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.38 | |
| I0409 03:05:39.993501 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 03:05:39.993513 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.82 | |
| I0409 03:05:39.993530 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.67346 (* 0.3 = 0.502038 loss) | |
| I0409 03:05:39.993546 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.494145 (* 0.3 = 0.148243 loss) | |
| I0409 03:05:39.993559 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.92 | |
| I0409 03:05:39.993572 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 03:05:39.993584 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 03:05:39.993599 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.372674 (* 1 = 0.372674 loss) | |
| I0409 03:05:39.993613 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.126909 (* 1 = 0.126909 loss) | |
| I0409 03:05:39.993626 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 03:05:39.993638 12249 solver.cpp:245] Train net output #16: total_confidence = 0.293725 | |
| I0409 03:05:39.993652 12249 sgd_solver.cpp:106] Iteration 101500, lr = 0.00855 | |
| I0409 03:11:13.372665 12249 solver.cpp:229] Iteration 102000, loss = 2.71757 | |
| I0409 03:11:13.372985 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.235294 | |
| I0409 03:11:13.373008 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 03:11:13.373020 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.529412 | |
| I0409 03:11:13.373036 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.50669 (* 0.3 = 0.752008 loss) | |
| I0409 03:11:13.373051 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.748485 (* 0.3 = 0.224545 loss) | |
| I0409 03:11:13.373064 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.411765 | |
| I0409 03:11:13.373076 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 03:11:13.373093 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.764706 | |
| I0409 03:11:13.373123 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.84712 (* 0.3 = 0.554137 loss) | |
| I0409 03:11:13.373142 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.567265 (* 0.3 = 0.17018 loss) | |
| I0409 03:11:13.373154 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.745098 | |
| I0409 03:11:13.373167 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 03:11:13.373178 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.921569 | |
| I0409 03:11:13.373193 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.938041 (* 1 = 0.938041 loss) | |
| I0409 03:11:13.373208 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.286184 (* 1 = 0.286184 loss) | |
| I0409 03:11:13.373224 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 03:11:13.373237 12249 solver.cpp:245] Train net output #16: total_confidence = 0.246338 | |
| I0409 03:11:13.373250 12249 sgd_solver.cpp:106] Iteration 102000, lr = 0.00854286 | |
| I0409 03:16:46.743373 12249 solver.cpp:229] Iteration 102500, loss = 2.77049 | |
| I0409 03:16:46.743486 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.254902 | |
| I0409 03:16:46.743506 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0409 03:16:46.743520 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.470588 | |
| I0409 03:16:46.743536 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.61965 (* 0.3 = 0.785895 loss) | |
| I0409 03:16:46.743551 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.830396 (* 0.3 = 0.249119 loss) | |
| I0409 03:16:46.743563 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.372549 | |
| I0409 03:16:46.743577 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0409 03:16:46.743588 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.686275 | |
| I0409 03:16:46.743602 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.00523 (* 0.3 = 0.601568 loss) | |
| I0409 03:16:46.743618 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.618849 (* 0.3 = 0.185655 loss) | |
| I0409 03:16:46.743629 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.607843 | |
| I0409 03:16:46.743641 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.875 | |
| I0409 03:16:46.743654 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.882353 | |
| I0409 03:16:46.743669 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.985973 (* 1 = 0.985973 loss) | |
| I0409 03:16:46.743682 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.330088 (* 1 = 0.330088 loss) | |
| I0409 03:16:46.743695 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 03:16:46.743706 12249 solver.cpp:245] Train net output #16: total_confidence = 0.140101 | |
| I0409 03:16:46.743719 12249 sgd_solver.cpp:106] Iteration 102500, lr = 0.00853571 | |
| I0409 03:22:20.100725 12249 solver.cpp:229] Iteration 103000, loss = 2.74896 | |
| I0409 03:22:20.101066 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3125 | |
| I0409 03:22:20.101099 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 03:22:20.101124 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.625 | |
| I0409 03:22:20.101153 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.17889 (* 0.3 = 0.653667 loss) | |
| I0409 03:22:20.101181 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.664792 (* 0.3 = 0.199438 loss) | |
| I0409 03:22:20.101203 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.354167 | |
| I0409 03:22:20.101227 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 03:22:20.101249 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.729167 | |
| I0409 03:22:20.101275 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.7197 (* 0.3 = 0.51591 loss) | |
| I0409 03:22:20.101301 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.502087 (* 0.3 = 0.150626 loss) | |
| I0409 03:22:20.101323 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.8125 | |
| I0409 03:22:20.101347 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 03:22:20.101369 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9375 | |
| I0409 03:22:20.101397 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.591239 (* 1 = 0.591239 loss) | |
| I0409 03:22:20.101423 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.177221 (* 1 = 0.177221 loss) | |
| I0409 03:22:20.101446 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 03:22:20.101467 12249 solver.cpp:245] Train net output #16: total_confidence = 0.296052 | |
| I0409 03:22:20.101491 12249 sgd_solver.cpp:106] Iteration 103000, lr = 0.00852857 | |
| I0409 03:27:53.470608 12249 solver.cpp:229] Iteration 103500, loss = 2.71607 | |
| I0409 03:27:53.470722 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.478261 | |
| I0409 03:27:53.470741 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 03:27:53.470757 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.673913 | |
| I0409 03:27:53.470774 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.87482 (* 0.3 = 0.562445 loss) | |
| I0409 03:27:53.470789 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.579948 (* 0.3 = 0.173985 loss) | |
| I0409 03:27:53.470801 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.543478 | |
| I0409 03:27:53.470814 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 03:27:53.470825 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.847826 | |
| I0409 03:27:53.470839 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.29754 (* 0.3 = 0.389262 loss) | |
| I0409 03:27:53.470854 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.396448 (* 0.3 = 0.118934 loss) | |
| I0409 03:27:53.470865 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.913043 | |
| I0409 03:27:53.470878 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 03:27:53.470891 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 03:27:53.470906 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.235279 (* 1 = 0.235279 loss) | |
| I0409 03:27:53.470919 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0739727 (* 1 = 0.0739727 loss) | |
| I0409 03:27:53.470932 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 03:27:53.470944 12249 solver.cpp:245] Train net output #16: total_confidence = 0.472802 | |
| I0409 03:27:53.470958 12249 sgd_solver.cpp:106] Iteration 103500, lr = 0.00852143 | |
| I0409 03:33:26.844339 12249 solver.cpp:229] Iteration 104000, loss = 2.74515 | |
| I0409 03:33:26.844715 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0409 03:33:26.844737 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0409 03:33:26.844753 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.538462 | |
| I0409 03:33:26.844770 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.34916 (* 0.3 = 0.704749 loss) | |
| I0409 03:33:26.844785 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.736475 (* 0.3 = 0.220943 loss) | |
| I0409 03:33:26.844799 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 03:33:26.844810 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 03:33:26.844823 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.673077 | |
| I0409 03:33:26.844837 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.73071 (* 0.3 = 0.519212 loss) | |
| I0409 03:33:26.844851 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.525049 (* 0.3 = 0.157515 loss) | |
| I0409 03:33:26.844863 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.884615 | |
| I0409 03:33:26.844876 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 03:33:26.844888 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.961538 | |
| I0409 03:33:26.844902 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.385275 (* 1 = 0.385275 loss) | |
| I0409 03:33:26.844916 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.117054 (* 1 = 0.117054 loss) | |
| I0409 03:33:26.844929 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 03:33:26.844941 12249 solver.cpp:245] Train net output #16: total_confidence = 0.384584 | |
| I0409 03:33:26.844954 12249 sgd_solver.cpp:106] Iteration 104000, lr = 0.00851429 | |
| I0409 03:39:00.217986 12249 solver.cpp:229] Iteration 104500, loss = 2.78757 | |
| I0409 03:39:00.218117 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.363636 | |
| I0409 03:39:00.218137 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 03:39:00.218152 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.636364 | |
| I0409 03:39:00.218168 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.29529 (* 0.3 = 0.688587 loss) | |
| I0409 03:39:00.218183 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.631543 (* 0.3 = 0.189463 loss) | |
| I0409 03:39:00.218196 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.477273 | |
| I0409 03:39:00.218209 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 03:39:00.218220 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.772727 | |
| I0409 03:39:00.218237 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.67317 (* 0.3 = 0.501951 loss) | |
| I0409 03:39:00.218252 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.520096 (* 0.3 = 0.156029 loss) | |
| I0409 03:39:00.218266 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.840909 | |
| I0409 03:39:00.218277 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 03:39:00.218289 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977273 | |
| I0409 03:39:00.218303 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.525933 (* 1 = 0.525933 loss) | |
| I0409 03:39:00.218317 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.150843 (* 1 = 0.150843 loss) | |
| I0409 03:39:00.218330 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 03:39:00.218343 12249 solver.cpp:245] Train net output #16: total_confidence = 0.180205 | |
| I0409 03:39:00.218356 12249 sgd_solver.cpp:106] Iteration 104500, lr = 0.00850714 | |
| I0409 03:44:33.189661 12249 solver.cpp:338] Iteration 105000, Testing net (#0) | |
| I0409 03:45:14.440274 12249 solver.cpp:393] Test loss: 2.74817 | |
| I0409 03:45:14.440388 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.330407 | |
| I0409 03:45:14.440407 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.81941 | |
| I0409 03:45:14.440420 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.599002 | |
| I0409 03:45:14.440436 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.32719 (* 0.3 = 0.698158 loss) | |
| I0409 03:45:14.440451 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.641159 (* 0.3 = 0.192348 loss) | |
| I0409 03:45:14.440464 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.566252 | |
| I0409 03:45:14.440476 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.882548 | |
| I0409 03:45:14.440487 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.817479 | |
| I0409 03:45:14.440501 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.52617 (* 0.3 = 0.45785 loss) | |
| I0409 03:45:14.440527 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.413966 (* 0.3 = 0.12419 loss) | |
| I0409 03:45:14.440542 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.757564 | |
| I0409 03:45:14.440554 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.936001 | |
| I0409 03:45:14.440565 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.883055 | |
| I0409 03:45:14.440579 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 1.00656 (* 1 = 1.00656 loss) | |
| I0409 03:45:14.440593 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.269067 (* 1 = 0.269067 loss) | |
| I0409 03:45:14.440605 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.395 | |
| I0409 03:45:14.440618 12249 solver.cpp:406] Test net output #16: total_confidence = 0.337421 | |
| I0409 03:45:14.818361 12249 solver.cpp:229] Iteration 105000, loss = 2.72172 | |
| I0409 03:45:14.818400 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.47619 | |
| I0409 03:45:14.818418 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 03:45:14.818430 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.690476 | |
| I0409 03:45:14.818446 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.83801 (* 0.3 = 0.551403 loss) | |
| I0409 03:45:14.818461 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.5524 (* 0.3 = 0.16572 loss) | |
| I0409 03:45:14.818473 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.642857 | |
| I0409 03:45:14.818485 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 03:45:14.818497 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.857143 | |
| I0409 03:45:14.818512 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.33707 (* 0.3 = 0.401122 loss) | |
| I0409 03:45:14.818526 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.431256 (* 0.3 = 0.129377 loss) | |
| I0409 03:45:14.818538 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.785714 | |
| I0409 03:45:14.818550 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 03:45:14.818562 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.928571 | |
| I0409 03:45:14.818577 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.649741 (* 1 = 0.649741 loss) | |
| I0409 03:45:14.818590 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.171255 (* 1 = 0.171255 loss) | |
| I0409 03:45:14.818603 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 03:45:14.818615 12249 solver.cpp:245] Train net output #16: total_confidence = 0.391635 | |
| I0409 03:45:14.818629 12249 sgd_solver.cpp:106] Iteration 105000, lr = 0.0085 | |
| I0409 03:50:48.305531 12249 solver.cpp:229] Iteration 105500, loss = 2.67123 | |
| I0409 03:50:48.305812 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.390244 | |
| I0409 03:50:48.305838 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 03:50:48.305852 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.512195 | |
| I0409 03:50:48.305869 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.50148 (* 0.3 = 0.750444 loss) | |
| I0409 03:50:48.305891 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.682568 (* 0.3 = 0.20477 loss) | |
| I0409 03:50:48.305904 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.512195 | |
| I0409 03:50:48.305917 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 03:50:48.305928 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.707317 | |
| I0409 03:50:48.305943 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.84592 (* 0.3 = 0.553775 loss) | |
| I0409 03:50:48.305958 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.517725 (* 0.3 = 0.155318 loss) | |
| I0409 03:50:48.305970 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.731707 | |
| I0409 03:50:48.305982 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 03:50:48.305994 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.951219 | |
| I0409 03:50:48.306008 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.668641 (* 1 = 0.668641 loss) | |
| I0409 03:50:48.306023 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.183783 (* 1 = 0.183783 loss) | |
| I0409 03:50:48.306035 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 03:50:48.306048 12249 solver.cpp:245] Train net output #16: total_confidence = 0.23782 | |
| I0409 03:50:48.306062 12249 sgd_solver.cpp:106] Iteration 105500, lr = 0.00849286 | |
| I0409 03:56:21.655602 12249 solver.cpp:229] Iteration 106000, loss = 2.7238 | |
| I0409 03:56:21.655704 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.342105 | |
| I0409 03:56:21.655722 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 03:56:21.655735 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.684211 | |
| I0409 03:56:21.655750 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.07748 (* 0.3 = 0.623245 loss) | |
| I0409 03:56:21.655766 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.617942 (* 0.3 = 0.185383 loss) | |
| I0409 03:56:21.655778 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.526316 | |
| I0409 03:56:21.655791 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 03:56:21.655802 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.763158 | |
| I0409 03:56:21.655817 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52879 (* 0.3 = 0.458637 loss) | |
| I0409 03:56:21.655832 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.571869 (* 0.3 = 0.171561 loss) | |
| I0409 03:56:21.655843 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.842105 | |
| I0409 03:56:21.655856 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 03:56:21.655869 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.868421 | |
| I0409 03:56:21.655882 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.44315 (* 1 = 1.44315 loss) | |
| I0409 03:56:21.655897 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.33347 (* 1 = 0.33347 loss) | |
| I0409 03:56:21.655910 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 03:56:21.655921 12249 solver.cpp:245] Train net output #16: total_confidence = 0.456408 | |
| I0409 03:56:21.655936 12249 sgd_solver.cpp:106] Iteration 106000, lr = 0.00848571 | |
| I0409 04:01:55.029731 12249 solver.cpp:229] Iteration 106500, loss = 2.70752 | |
| I0409 04:01:55.030062 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.185185 | |
| I0409 04:01:55.030084 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.738636 | |
| I0409 04:01:55.030097 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.37037 | |
| I0409 04:01:55.030113 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.18999 (* 0.3 = 0.956996 loss) | |
| I0409 04:01:55.030129 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.03141 (* 0.3 = 0.309424 loss) | |
| I0409 04:01:55.030141 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.37037 | |
| I0409 04:01:55.030154 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0409 04:01:55.030165 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.62963 | |
| I0409 04:01:55.030179 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.28864 (* 0.3 = 0.686593 loss) | |
| I0409 04:01:55.030194 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.723983 (* 0.3 = 0.217195 loss) | |
| I0409 04:01:55.030206 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.62963 | |
| I0409 04:01:55.030220 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0409 04:01:55.030231 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.87037 | |
| I0409 04:01:55.030246 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.39631 (* 1 = 1.39631 loss) | |
| I0409 04:01:55.030259 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.434571 (* 1 = 0.434571 loss) | |
| I0409 04:01:55.030272 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 04:01:55.030283 12249 solver.cpp:245] Train net output #16: total_confidence = 0.11214 | |
| I0409 04:01:55.030298 12249 sgd_solver.cpp:106] Iteration 106500, lr = 0.00847857 | |
| I0409 04:07:28.394511 12249 solver.cpp:229] Iteration 107000, loss = 2.68768 | |
| I0409 04:07:28.394632 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.348837 | |
| I0409 04:07:28.394652 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 04:07:28.394665 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.627907 | |
| I0409 04:07:28.394682 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.16244 (* 0.3 = 0.648731 loss) | |
| I0409 04:07:28.394697 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.625195 (* 0.3 = 0.187558 loss) | |
| I0409 04:07:28.394711 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.511628 | |
| I0409 04:07:28.394722 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 04:07:28.394734 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.837209 | |
| I0409 04:07:28.394752 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.5257 (* 0.3 = 0.457711 loss) | |
| I0409 04:07:28.394767 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.46315 (* 0.3 = 0.138945 loss) | |
| I0409 04:07:28.394779 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.860465 | |
| I0409 04:07:28.394791 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 04:07:28.394804 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0409 04:07:28.394819 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.561866 (* 1 = 0.561866 loss) | |
| I0409 04:07:28.394832 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.171024 (* 1 = 0.171024 loss) | |
| I0409 04:07:28.394845 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 04:07:28.394857 12249 solver.cpp:245] Train net output #16: total_confidence = 0.3148 | |
| I0409 04:07:28.394872 12249 sgd_solver.cpp:106] Iteration 107000, lr = 0.00847143 | |
| I0409 04:13:01.787242 12249 solver.cpp:229] Iteration 107500, loss = 2.69857 | |
| I0409 04:13:01.787536 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.416667 | |
| I0409 04:13:01.787557 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 04:13:01.787570 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.708333 | |
| I0409 04:13:01.787587 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.11725 (* 0.3 = 0.635176 loss) | |
| I0409 04:13:01.787602 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.626035 (* 0.3 = 0.187811 loss) | |
| I0409 04:13:01.787616 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.541667 | |
| I0409 04:13:01.787627 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 04:13:01.787639 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.770833 | |
| I0409 04:13:01.787653 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.54964 (* 0.3 = 0.464892 loss) | |
| I0409 04:13:01.787667 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.474538 (* 0.3 = 0.142361 loss) | |
| I0409 04:13:01.787680 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.916667 | |
| I0409 04:13:01.787693 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 04:13:01.787704 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.958333 | |
| I0409 04:13:01.787719 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.51343 (* 1 = 0.51343 loss) | |
| I0409 04:13:01.787732 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.174248 (* 1 = 0.174248 loss) | |
| I0409 04:13:01.787747 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 04:13:01.787760 12249 solver.cpp:245] Train net output #16: total_confidence = 0.359011 | |
| I0409 04:13:01.787775 12249 sgd_solver.cpp:106] Iteration 107500, lr = 0.00846429 | |
| I0409 04:18:35.135864 12249 solver.cpp:229] Iteration 108000, loss = 2.70461 | |
| I0409 04:18:35.136003 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.269231 | |
| I0409 04:18:35.136024 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 04:18:35.136037 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.538462 | |
| I0409 04:18:35.136054 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.40879 (* 0.3 = 0.722638 loss) | |
| I0409 04:18:35.136070 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.74 (* 0.3 = 0.222 loss) | |
| I0409 04:18:35.136083 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.403846 | |
| I0409 04:18:35.136096 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 04:18:35.136107 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.692308 | |
| I0409 04:18:35.136121 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.75549 (* 0.3 = 0.526647 loss) | |
| I0409 04:18:35.136135 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.529006 (* 0.3 = 0.158702 loss) | |
| I0409 04:18:35.136147 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826923 | |
| I0409 04:18:35.136159 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 04:18:35.136171 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.942308 | |
| I0409 04:18:35.136185 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.561139 (* 1 = 0.561139 loss) | |
| I0409 04:18:35.136200 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.237048 (* 1 = 0.237048 loss) | |
| I0409 04:18:35.136212 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 04:18:35.136224 12249 solver.cpp:245] Train net output #16: total_confidence = 0.224325 | |
| I0409 04:18:35.136240 12249 sgd_solver.cpp:106] Iteration 108000, lr = 0.00845714 | |
| I0409 04:24:08.508927 12249 solver.cpp:229] Iteration 108500, loss = 2.71684 | |
| I0409 04:24:08.509259 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.291667 | |
| I0409 04:24:08.509281 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 04:24:08.509295 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.625 | |
| I0409 04:24:08.509311 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.14166 (* 0.3 = 0.642499 loss) | |
| I0409 04:24:08.509326 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.625065 (* 0.3 = 0.18752 loss) | |
| I0409 04:24:08.509341 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.479167 | |
| I0409 04:24:08.509352 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 04:24:08.509364 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.6875 | |
| I0409 04:24:08.509378 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.76117 (* 0.3 = 0.52835 loss) | |
| I0409 04:24:08.509392 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.548789 (* 0.3 = 0.164637 loss) | |
| I0409 04:24:08.509404 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.708333 | |
| I0409 04:24:08.509418 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 04:24:08.509429 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.895833 | |
| I0409 04:24:08.509443 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.846494 (* 1 = 0.846494 loss) | |
| I0409 04:24:08.509459 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.249335 (* 1 = 0.249335 loss) | |
| I0409 04:24:08.509471 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 04:24:08.509484 12249 solver.cpp:245] Train net output #16: total_confidence = 0.289246 | |
| I0409 04:24:08.509497 12249 sgd_solver.cpp:106] Iteration 108500, lr = 0.00845 | |
| I0409 04:29:42.549311 12249 solver.cpp:229] Iteration 109000, loss = 2.65821 | |
| I0409 04:29:42.549460 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.388889 | |
| I0409 04:29:42.549480 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 04:29:42.549494 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.666667 | |
| I0409 04:29:42.549510 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.9638 (* 0.3 = 0.58914 loss) | |
| I0409 04:29:42.549525 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.553321 (* 0.3 = 0.165996 loss) | |
| I0409 04:29:42.549537 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.583333 | |
| I0409 04:29:42.549551 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 04:29:42.549562 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.833333 | |
| I0409 04:29:42.549576 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52361 (* 0.3 = 0.457084 loss) | |
| I0409 04:29:42.549590 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.494581 (* 0.3 = 0.148374 loss) | |
| I0409 04:29:42.549602 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.888889 | |
| I0409 04:29:42.549614 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 04:29:42.549626 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.972222 | |
| I0409 04:29:42.549641 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.413921 (* 1 = 0.413921 loss) | |
| I0409 04:29:42.549655 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.129267 (* 1 = 0.129267 loss) | |
| I0409 04:29:42.549667 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 04:29:42.549680 12249 solver.cpp:245] Train net output #16: total_confidence = 0.407029 | |
| I0409 04:29:42.549695 12249 sgd_solver.cpp:106] Iteration 109000, lr = 0.00844286 | |
| I0409 04:35:15.908308 12249 solver.cpp:229] Iteration 109500, loss = 2.75269 | |
| I0409 04:35:15.908650 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.232558 | |
| I0409 04:35:15.908671 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 04:35:15.908685 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.581395 | |
| I0409 04:35:15.908701 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.51599 (* 0.3 = 0.754796 loss) | |
| I0409 04:35:15.908717 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.686869 (* 0.3 = 0.206061 loss) | |
| I0409 04:35:15.908730 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.418605 | |
| I0409 04:35:15.908745 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 04:35:15.908757 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.767442 | |
| I0409 04:35:15.908771 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.756 (* 0.3 = 0.5268 loss) | |
| I0409 04:35:15.908787 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.479939 (* 0.3 = 0.143982 loss) | |
| I0409 04:35:15.908799 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.767442 | |
| I0409 04:35:15.908812 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 04:35:15.908823 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0409 04:35:15.908838 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.915436 (* 1 = 0.915436 loss) | |
| I0409 04:35:15.908851 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.272046 (* 1 = 0.272046 loss) | |
| I0409 04:35:15.908864 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 04:35:15.908875 12249 solver.cpp:245] Train net output #16: total_confidence = 0.181166 | |
| I0409 04:35:15.908890 12249 sgd_solver.cpp:106] Iteration 109500, lr = 0.00843571 | |
| I0409 04:40:48.894635 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_110000.caffemodel | |
| I0409 04:40:49.332358 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_110000.solverstate | |
| I0409 04:40:49.571470 12249 solver.cpp:338] Iteration 110000, Testing net (#0) | |
| I0409 04:41:30.531529 12249 solver.cpp:393] Test loss: 2.40213 | |
| I0409 04:41:30.531661 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.414106 | |
| I0409 04:41:30.531680 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.839911 | |
| I0409 04:41:30.531693 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.729206 | |
| I0409 04:41:30.531709 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.86584 (* 0.3 = 0.559753 loss) | |
| I0409 04:41:30.531724 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.524951 (* 0.3 = 0.157485 loss) | |
| I0409 04:41:30.531736 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.604912 | |
| I0409 04:41:30.531752 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.886867 | |
| I0409 04:41:30.531764 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.864748 | |
| I0409 04:41:30.531779 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.37329 (* 0.3 = 0.411987 loss) | |
| I0409 04:41:30.531792 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.389873 (* 0.3 = 0.116962 loss) | |
| I0409 04:41:30.531805 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.786191 | |
| I0409 04:41:30.531816 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.945546 | |
| I0409 04:41:30.531827 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.901554 | |
| I0409 04:41:30.531841 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.91754 (* 1 = 0.91754 loss) | |
| I0409 04:41:30.531855 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.238407 (* 1 = 0.238407 loss) | |
| I0409 04:41:30.531867 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.468 | |
| I0409 04:41:30.531878 12249 solver.cpp:406] Test net output #16: total_confidence = 0.427525 | |
| I0409 04:41:30.903426 12249 solver.cpp:229] Iteration 110000, loss = 2.61311 | |
| I0409 04:41:30.903487 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.288889 | |
| I0409 04:41:30.903504 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 04:41:30.903517 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.533333 | |
| I0409 04:41:30.903533 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.7857 (* 0.3 = 0.835711 loss) | |
| I0409 04:41:30.903548 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.854579 (* 0.3 = 0.256374 loss) | |
| I0409 04:41:30.903561 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4 | |
| I0409 04:41:30.903574 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 04:41:30.903586 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.688889 | |
| I0409 04:41:30.903600 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.17935 (* 0.3 = 0.653804 loss) | |
| I0409 04:41:30.903614 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.658121 (* 0.3 = 0.197436 loss) | |
| I0409 04:41:30.903626 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.688889 | |
| I0409 04:41:30.903640 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 04:41:30.903651 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.866667 | |
| I0409 04:41:30.903666 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.05531 (* 1 = 1.05531 loss) | |
| I0409 04:41:30.903679 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.328299 (* 1 = 0.328299 loss) | |
| I0409 04:41:30.903692 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 04:41:30.903704 12249 solver.cpp:245] Train net output #16: total_confidence = 0.317682 | |
| I0409 04:41:30.903719 12249 sgd_solver.cpp:106] Iteration 110000, lr = 0.00842857 | |
| I0409 04:47:04.327121 12249 solver.cpp:229] Iteration 110500, loss = 2.66001 | |
| I0409 04:47:04.327265 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.395833 | |
| I0409 04:47:04.327285 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 04:47:04.327298 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.645833 | |
| I0409 04:47:04.327314 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.06597 (* 0.3 = 0.619791 loss) | |
| I0409 04:47:04.327329 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.619191 (* 0.3 = 0.185757 loss) | |
| I0409 04:47:04.327342 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.520833 | |
| I0409 04:47:04.327356 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 04:47:04.327368 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.770833 | |
| I0409 04:47:04.327383 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.63063 (* 0.3 = 0.489188 loss) | |
| I0409 04:47:04.327396 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.489471 (* 0.3 = 0.146841 loss) | |
| I0409 04:47:04.327409 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.770833 | |
| I0409 04:47:04.327421 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 04:47:04.327433 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.895833 | |
| I0409 04:47:04.327450 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.81771 (* 1 = 0.81771 loss) | |
| I0409 04:47:04.327464 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.228639 (* 1 = 0.228639 loss) | |
| I0409 04:47:04.327476 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 04:47:04.327488 12249 solver.cpp:245] Train net output #16: total_confidence = 0.342129 | |
| I0409 04:47:04.327502 12249 sgd_solver.cpp:106] Iteration 110500, lr = 0.00842143 | |
| I0409 04:52:37.693267 12249 solver.cpp:229] Iteration 111000, loss = 2.60661 | |
| I0409 04:52:37.693577 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.27451 | |
| I0409 04:52:37.693598 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0409 04:52:37.693611 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.490196 | |
| I0409 04:52:37.693629 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.67312 (* 0.3 = 0.801937 loss) | |
| I0409 04:52:37.693645 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.870185 (* 0.3 = 0.261055 loss) | |
| I0409 04:52:37.693656 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.470588 | |
| I0409 04:52:37.693670 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 04:52:37.693681 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72549 | |
| I0409 04:52:37.693696 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.94329 (* 0.3 = 0.582986 loss) | |
| I0409 04:52:37.693711 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.636981 (* 0.3 = 0.191094 loss) | |
| I0409 04:52:37.693722 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.588235 | |
| I0409 04:52:37.693734 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.863636 | |
| I0409 04:52:37.693749 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.862745 | |
| I0409 04:52:37.693764 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.12632 (* 1 = 1.12632 loss) | |
| I0409 04:52:37.693778 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.388524 (* 1 = 0.388524 loss) | |
| I0409 04:52:37.693791 12249 solver.cpp:245] Train net output #15: total_accuracy = 0 | |
| I0409 04:52:37.693804 12249 solver.cpp:245] Train net output #16: total_confidence = 0.20886 | |
| I0409 04:52:37.693817 12249 sgd_solver.cpp:106] Iteration 111000, lr = 0.00841429 | |
| I0409 04:58:11.061836 12249 solver.cpp:229] Iteration 111500, loss = 2.6425 | |
| I0409 04:58:11.061924 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.288889 | |
| I0409 04:58:11.061944 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 04:58:11.061956 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.555556 | |
| I0409 04:58:11.061972 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.43251 (* 0.3 = 0.729754 loss) | |
| I0409 04:58:11.061988 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.716559 (* 0.3 = 0.214968 loss) | |
| I0409 04:58:11.062000 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.577778 | |
| I0409 04:58:11.062012 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 04:58:11.062024 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.822222 | |
| I0409 04:58:11.062039 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.54559 (* 0.3 = 0.463676 loss) | |
| I0409 04:58:11.062053 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.454343 (* 0.3 = 0.136303 loss) | |
| I0409 04:58:11.062065 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.911111 | |
| I0409 04:58:11.062078 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 04:58:11.062089 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.933333 | |
| I0409 04:58:11.062103 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.589383 (* 1 = 0.589383 loss) | |
| I0409 04:58:11.062119 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.158338 (* 1 = 0.158338 loss) | |
| I0409 04:58:11.062130 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 04:58:11.062142 12249 solver.cpp:245] Train net output #16: total_confidence = 0.529967 | |
| I0409 04:58:11.062157 12249 sgd_solver.cpp:106] Iteration 111500, lr = 0.00840714 | |
| I0409 05:03:44.441897 12249 solver.cpp:229] Iteration 112000, loss = 2.64588 | |
| I0409 05:03:44.442188 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.355556 | |
| I0409 05:03:44.442208 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 05:03:44.442221 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.644444 | |
| I0409 05:03:44.442239 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.2852 (* 0.3 = 0.685559 loss) | |
| I0409 05:03:44.442253 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.708562 (* 0.3 = 0.212569 loss) | |
| I0409 05:03:44.442266 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.555556 | |
| I0409 05:03:44.442279 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 05:03:44.442291 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.777778 | |
| I0409 05:03:44.442304 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.62828 (* 0.3 = 0.488483 loss) | |
| I0409 05:03:44.442318 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.539649 (* 0.3 = 0.161895 loss) | |
| I0409 05:03:44.442332 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.844444 | |
| I0409 05:03:44.442343 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 05:03:44.442355 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.955556 | |
| I0409 05:03:44.442369 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.582862 (* 1 = 0.582862 loss) | |
| I0409 05:03:44.442384 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.170433 (* 1 = 0.170433 loss) | |
| I0409 05:03:44.442396 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 05:03:44.442409 12249 solver.cpp:245] Train net output #16: total_confidence = 0.235643 | |
| I0409 05:03:44.442422 12249 sgd_solver.cpp:106] Iteration 112000, lr = 0.0084 | |
| I0409 05:09:17.803278 12249 solver.cpp:229] Iteration 112500, loss = 2.70788 | |
| I0409 05:09:17.803407 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0409 05:09:17.803427 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 05:09:17.803442 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.604167 | |
| I0409 05:09:17.803458 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.30139 (* 0.3 = 0.690416 loss) | |
| I0409 05:09:17.803473 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.730051 (* 0.3 = 0.219015 loss) | |
| I0409 05:09:17.803486 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.541667 | |
| I0409 05:09:17.803499 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 05:09:17.803511 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.770833 | |
| I0409 05:09:17.803525 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.51592 (* 0.3 = 0.454776 loss) | |
| I0409 05:09:17.803540 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.485536 (* 0.3 = 0.145661 loss) | |
| I0409 05:09:17.803552 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.854167 | |
| I0409 05:09:17.803565 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 05:09:17.803577 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9375 | |
| I0409 05:09:17.803592 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.726794 (* 1 = 0.726794 loss) | |
| I0409 05:09:17.803606 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.21586 (* 1 = 0.21586 loss) | |
| I0409 05:09:17.803618 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 05:09:17.803630 12249 solver.cpp:245] Train net output #16: total_confidence = 0.302705 | |
| I0409 05:09:17.803645 12249 sgd_solver.cpp:106] Iteration 112500, lr = 0.00839286 | |
| I0409 05:14:51.171010 12249 solver.cpp:229] Iteration 113000, loss = 2.62513 | |
| I0409 05:14:51.171306 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.188679 | |
| I0409 05:14:51.171329 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.75 | |
| I0409 05:14:51.171342 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.415094 | |
| I0409 05:14:51.171358 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.31856 (* 0.3 = 0.995568 loss) | |
| I0409 05:14:51.171375 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 1.03154 (* 0.3 = 0.309463 loss) | |
| I0409 05:14:51.171386 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.320755 | |
| I0409 05:14:51.171399 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.789773 | |
| I0409 05:14:51.171411 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.490566 | |
| I0409 05:14:51.171425 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.80231 (* 0.3 = 0.840694 loss) | |
| I0409 05:14:51.171439 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.871357 (* 0.3 = 0.261407 loss) | |
| I0409 05:14:51.171452 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.528302 | |
| I0409 05:14:51.171465 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.852273 | |
| I0409 05:14:51.171478 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.603774 | |
| I0409 05:14:51.171491 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.13191 (* 1 = 2.13191 loss) | |
| I0409 05:14:51.171505 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.657888 (* 1 = 0.657888 loss) | |
| I0409 05:14:51.171519 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 05:14:51.171530 12249 solver.cpp:245] Train net output #16: total_confidence = 0.220002 | |
| I0409 05:14:51.171545 12249 sgd_solver.cpp:106] Iteration 113000, lr = 0.00838571 | |
| I0409 05:20:24.554216 12249 solver.cpp:229] Iteration 113500, loss = 2.65347 | |
| I0409 05:20:24.554339 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.333333 | |
| I0409 05:20:24.554359 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 05:20:24.554373 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.622222 | |
| I0409 05:20:24.554389 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.6574 (* 0.3 = 0.797219 loss) | |
| I0409 05:20:24.554404 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.785619 (* 0.3 = 0.235686 loss) | |
| I0409 05:20:24.554417 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.377778 | |
| I0409 05:20:24.554430 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 05:20:24.554442 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.6 | |
| I0409 05:20:24.554456 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.03341 (* 0.3 = 0.610023 loss) | |
| I0409 05:20:24.554471 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.589958 (* 0.3 = 0.176987 loss) | |
| I0409 05:20:24.554483 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.622222 | |
| I0409 05:20:24.554496 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0409 05:20:24.554507 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.844444 | |
| I0409 05:20:24.554522 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.33337 (* 1 = 1.33337 loss) | |
| I0409 05:20:24.554535 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.361866 (* 1 = 0.361866 loss) | |
| I0409 05:20:24.554548 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 05:20:24.554559 12249 solver.cpp:245] Train net output #16: total_confidence = 0.430326 | |
| I0409 05:20:24.554574 12249 sgd_solver.cpp:106] Iteration 113500, lr = 0.00837857 | |
| I0409 05:25:57.916656 12249 solver.cpp:229] Iteration 114000, loss = 2.59866 | |
| I0409 05:25:57.916972 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.348837 | |
| I0409 05:25:57.916995 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 05:25:57.917007 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.488372 | |
| I0409 05:25:57.917024 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.49373 (* 0.3 = 0.74812 loss) | |
| I0409 05:25:57.917039 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.686297 (* 0.3 = 0.205889 loss) | |
| I0409 05:25:57.917052 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.44186 | |
| I0409 05:25:57.917064 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 05:25:57.917076 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.790698 | |
| I0409 05:25:57.917090 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.68256 (* 0.3 = 0.504768 loss) | |
| I0409 05:25:57.917105 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.516466 (* 0.3 = 0.15494 loss) | |
| I0409 05:25:57.917117 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.674419 | |
| I0409 05:25:57.917129 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 05:25:57.917141 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.906977 | |
| I0409 05:25:57.917156 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.835534 (* 1 = 0.835534 loss) | |
| I0409 05:25:57.917171 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.277481 (* 1 = 0.277481 loss) | |
| I0409 05:25:57.917183 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 05:25:57.917196 12249 solver.cpp:245] Train net output #16: total_confidence = 0.268015 | |
| I0409 05:25:57.917209 12249 sgd_solver.cpp:106] Iteration 114000, lr = 0.00837143 | |
| I0409 05:31:31.283607 12249 solver.cpp:229] Iteration 114500, loss = 2.57998 | |
| I0409 05:31:31.283859 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 05:31:31.283880 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 05:31:31.283893 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6 | |
| I0409 05:31:31.283910 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.04304 (* 0.3 = 0.612912 loss) | |
| I0409 05:31:31.283926 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.576116 (* 0.3 = 0.172835 loss) | |
| I0409 05:31:31.283938 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.666667 | |
| I0409 05:31:31.283951 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 05:31:31.283963 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.844444 | |
| I0409 05:31:31.283977 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.22396 (* 0.3 = 0.367187 loss) | |
| I0409 05:31:31.283993 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.363359 (* 0.3 = 0.109008 loss) | |
| I0409 05:31:31.284004 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.933333 | |
| I0409 05:31:31.284016 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 05:31:31.284029 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977778 | |
| I0409 05:31:31.284044 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.256894 (* 1 = 0.256894 loss) | |
| I0409 05:31:31.284059 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0720504 (* 1 = 0.0720504 loss) | |
| I0409 05:31:31.284071 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 05:31:31.284083 12249 solver.cpp:245] Train net output #16: total_confidence = 0.459295 | |
| I0409 05:31:31.284097 12249 sgd_solver.cpp:106] Iteration 114500, lr = 0.00836429 | |
| I0409 05:37:04.266458 12249 solver.cpp:338] Iteration 115000, Testing net (#0) | |
| I0409 05:37:45.596933 12249 solver.cpp:393] Test loss: 2.57863 | |
| I0409 05:37:45.597046 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.363978 | |
| I0409 05:37:45.597066 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.839046 | |
| I0409 05:37:45.597079 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.659029 | |
| I0409 05:37:45.597095 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 2.25259 (* 0.3 = 0.675776 loss) | |
| I0409 05:37:45.597111 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.578006 (* 0.3 = 0.173402 loss) | |
| I0409 05:37:45.597123 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.575355 | |
| I0409 05:37:45.597136 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.887049 | |
| I0409 05:37:45.597147 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.850998 | |
| I0409 05:37:45.597162 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.43872 (* 0.3 = 0.431617 loss) | |
| I0409 05:37:45.597175 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.387128 (* 0.3 = 0.116139 loss) | |
| I0409 05:37:45.597187 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.766218 | |
| I0409 05:37:45.597199 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.940683 | |
| I0409 05:37:45.597210 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.902077 | |
| I0409 05:37:45.597225 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.938187 (* 1 = 0.938187 loss) | |
| I0409 05:37:45.597239 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.243513 (* 1 = 0.243513 loss) | |
| I0409 05:37:45.597250 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.408 | |
| I0409 05:37:45.597261 12249 solver.cpp:406] Test net output #16: total_confidence = 0.382974 | |
| I0409 05:37:45.976344 12249 solver.cpp:229] Iteration 115000, loss = 2.59329 | |
| I0409 05:37:45.976404 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.275 | |
| I0409 05:37:45.976423 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 05:37:45.976436 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.625 | |
| I0409 05:37:45.976454 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.26109 (* 0.3 = 0.678327 loss) | |
| I0409 05:37:45.976469 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.684197 (* 0.3 = 0.205259 loss) | |
| I0409 05:37:45.976493 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 05:37:45.976510 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 05:37:45.976522 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.7 | |
| I0409 05:37:45.976536 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.7208 (* 0.3 = 0.516239 loss) | |
| I0409 05:37:45.976552 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.541828 (* 0.3 = 0.162548 loss) | |
| I0409 05:37:45.976563 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.7 | |
| I0409 05:37:45.976577 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 05:37:45.976588 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9 | |
| I0409 05:37:45.976603 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.02145 (* 1 = 1.02145 loss) | |
| I0409 05:37:45.976616 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.38738 (* 1 = 0.38738 loss) | |
| I0409 05:37:45.976629 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 05:37:45.976641 12249 solver.cpp:245] Train net output #16: total_confidence = 0.284379 | |
| I0409 05:37:45.976656 12249 sgd_solver.cpp:106] Iteration 115000, lr = 0.00835714 | |
| I0409 05:43:19.343511 12249 solver.cpp:229] Iteration 115500, loss = 2.58224 | |
| I0409 05:43:19.343828 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.277778 | |
| I0409 05:43:19.343852 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 05:43:19.343864 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.574074 | |
| I0409 05:43:19.343881 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.39226 (* 0.3 = 0.717679 loss) | |
| I0409 05:43:19.343896 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.742962 (* 0.3 = 0.222889 loss) | |
| I0409 05:43:19.343909 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.351852 | |
| I0409 05:43:19.343921 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.801136 | |
| I0409 05:43:19.343933 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.703704 | |
| I0409 05:43:19.343947 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.1377 (* 0.3 = 0.64131 loss) | |
| I0409 05:43:19.343961 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.658745 (* 0.3 = 0.197623 loss) | |
| I0409 05:43:19.343974 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.814815 | |
| I0409 05:43:19.343986 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 05:43:19.343998 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.888889 | |
| I0409 05:43:19.344012 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.909903 (* 1 = 0.909903 loss) | |
| I0409 05:43:19.344027 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.285447 (* 1 = 0.285447 loss) | |
| I0409 05:43:19.344039 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 05:43:19.344051 12249 solver.cpp:245] Train net output #16: total_confidence = 0.353419 | |
| I0409 05:43:19.344065 12249 sgd_solver.cpp:106] Iteration 115500, lr = 0.00835 | |
| I0409 05:48:52.722563 12249 solver.cpp:229] Iteration 116000, loss = 2.66869 | |
| I0409 05:48:52.722676 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.28 | |
| I0409 05:48:52.722694 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 05:48:52.722707 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6 | |
| I0409 05:48:52.722723 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.26088 (* 0.3 = 0.678265 loss) | |
| I0409 05:48:52.722739 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.688197 (* 0.3 = 0.206459 loss) | |
| I0409 05:48:52.722751 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.46 | |
| I0409 05:48:52.722764 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 05:48:52.722776 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.76 | |
| I0409 05:48:52.722790 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.71486 (* 0.3 = 0.514458 loss) | |
| I0409 05:48:52.722803 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.567069 (* 0.3 = 0.170121 loss) | |
| I0409 05:48:52.722816 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.82 | |
| I0409 05:48:52.722828 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 05:48:52.722841 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.88 | |
| I0409 05:48:52.722854 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.924655 (* 1 = 0.924655 loss) | |
| I0409 05:48:52.722868 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.34887 (* 1 = 0.34887 loss) | |
| I0409 05:48:52.722880 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 05:48:52.722892 12249 solver.cpp:245] Train net output #16: total_confidence = 0.241547 | |
| I0409 05:48:52.722906 12249 sgd_solver.cpp:106] Iteration 116000, lr = 0.00834286 | |
| I0409 05:54:26.091040 12249 solver.cpp:229] Iteration 116500, loss = 2.59315 | |
| I0409 05:54:26.091362 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.266667 | |
| I0409 05:54:26.091383 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 05:54:26.091397 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.488889 | |
| I0409 05:54:26.091413 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.28567 (* 0.3 = 0.685702 loss) | |
| I0409 05:54:26.091428 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.626359 (* 0.3 = 0.187908 loss) | |
| I0409 05:54:26.091441 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.444444 | |
| I0409 05:54:26.091455 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 05:54:26.091467 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.777778 | |
| I0409 05:54:26.091481 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.70863 (* 0.3 = 0.512588 loss) | |
| I0409 05:54:26.091495 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.46484 (* 0.3 = 0.139452 loss) | |
| I0409 05:54:26.091508 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.733333 | |
| I0409 05:54:26.091521 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 05:54:26.091532 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.911111 | |
| I0409 05:54:26.091547 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.934214 (* 1 = 0.934214 loss) | |
| I0409 05:54:26.091562 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.252488 (* 1 = 0.252488 loss) | |
| I0409 05:54:26.091573 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 05:54:26.091586 12249 solver.cpp:245] Train net output #16: total_confidence = 0.362665 | |
| I0409 05:54:26.091600 12249 sgd_solver.cpp:106] Iteration 116500, lr = 0.00833571 | |
| I0409 05:59:59.801210 12249 solver.cpp:229] Iteration 117000, loss = 2.58426 | |
| I0409 05:59:59.801358 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.255814 | |
| I0409 05:59:59.801385 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 05:59:59.801398 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.55814 | |
| I0409 05:59:59.801415 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.29355 (* 0.3 = 0.688065 loss) | |
| I0409 05:59:59.801430 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.681346 (* 0.3 = 0.204404 loss) | |
| I0409 05:59:59.801443 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.418605 | |
| I0409 05:59:59.801456 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 05:59:59.801467 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.767442 | |
| I0409 05:59:59.801481 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.80541 (* 0.3 = 0.541624 loss) | |
| I0409 05:59:59.801496 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.511877 (* 0.3 = 0.153563 loss) | |
| I0409 05:59:59.801507 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.790698 | |
| I0409 05:59:59.801520 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 05:59:59.801533 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.906977 | |
| I0409 05:59:59.801548 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.755646 (* 1 = 0.755646 loss) | |
| I0409 05:59:59.801561 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.217797 (* 1 = 0.217797 loss) | |
| I0409 05:59:59.801574 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 05:59:59.801586 12249 solver.cpp:245] Train net output #16: total_confidence = 0.271197 | |
| I0409 05:59:59.801600 12249 sgd_solver.cpp:106] Iteration 117000, lr = 0.00832857 | |
| I0409 06:05:33.165406 12249 solver.cpp:229] Iteration 117500, loss = 2.59464 | |
| I0409 06:05:33.165710 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.395833 | |
| I0409 06:05:33.165730 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 06:05:33.165745 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.604167 | |
| I0409 06:05:33.165763 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.11803 (* 0.3 = 0.635409 loss) | |
| I0409 06:05:33.165778 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.622613 (* 0.3 = 0.186784 loss) | |
| I0409 06:05:33.165791 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4375 | |
| I0409 06:05:33.165803 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 06:05:33.165817 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.770833 | |
| I0409 06:05:33.165830 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.47522 (* 0.3 = 0.442566 loss) | |
| I0409 06:05:33.165844 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.452359 (* 0.3 = 0.135708 loss) | |
| I0409 06:05:33.165858 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.770833 | |
| I0409 06:05:33.165869 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 06:05:33.165881 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9375 | |
| I0409 06:05:33.165896 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.68231 (* 1 = 0.68231 loss) | |
| I0409 06:05:33.165911 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.210273 (* 1 = 0.210273 loss) | |
| I0409 06:05:33.165923 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 06:05:33.165935 12249 solver.cpp:245] Train net output #16: total_confidence = 0.374987 | |
| I0409 06:05:33.165949 12249 sgd_solver.cpp:106] Iteration 117500, lr = 0.00832143 | |
| I0409 06:11:06.537881 12249 solver.cpp:229] Iteration 118000, loss = 2.57512 | |
| I0409 06:11:06.538103 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.444444 | |
| I0409 06:11:06.538122 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 06:11:06.538136 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.644444 | |
| I0409 06:11:06.538152 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.02952 (* 0.3 = 0.608855 loss) | |
| I0409 06:11:06.538167 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.596675 (* 0.3 = 0.179002 loss) | |
| I0409 06:11:06.538180 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.577778 | |
| I0409 06:11:06.538192 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 06:11:06.538204 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.888889 | |
| I0409 06:11:06.538218 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.36613 (* 0.3 = 0.409838 loss) | |
| I0409 06:11:06.538233 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.460707 (* 0.3 = 0.138212 loss) | |
| I0409 06:11:06.538245 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.866667 | |
| I0409 06:11:06.538257 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 06:11:06.538269 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 06:11:06.538283 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.345451 (* 1 = 0.345451 loss) | |
| I0409 06:11:06.538297 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.134642 (* 1 = 0.134642 loss) | |
| I0409 06:11:06.538310 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 06:11:06.538324 12249 solver.cpp:245] Train net output #16: total_confidence = 0.383356 | |
| I0409 06:11:06.538339 12249 sgd_solver.cpp:106] Iteration 118000, lr = 0.00831429 | |
| I0409 06:16:39.906896 12249 solver.cpp:229] Iteration 118500, loss = 2.60407 | |
| I0409 06:16:39.907052 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.434783 | |
| I0409 06:16:39.907071 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 06:16:39.907084 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.804348 | |
| I0409 06:16:39.907101 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.70922 (* 0.3 = 0.512766 loss) | |
| I0409 06:16:39.907116 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.54123 (* 0.3 = 0.162369 loss) | |
| I0409 06:16:39.907130 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.586957 | |
| I0409 06:16:39.907141 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 06:16:39.907153 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.891304 | |
| I0409 06:16:39.907167 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.19681 (* 0.3 = 0.359042 loss) | |
| I0409 06:16:39.907182 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.394906 (* 0.3 = 0.118472 loss) | |
| I0409 06:16:39.907194 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.782609 | |
| I0409 06:16:39.907207 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 06:16:39.907219 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 06:16:39.907234 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.635636 (* 1 = 0.635636 loss) | |
| I0409 06:16:39.907248 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.199847 (* 1 = 0.199847 loss) | |
| I0409 06:16:39.907261 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 06:16:39.907274 12249 solver.cpp:245] Train net output #16: total_confidence = 0.415608 | |
| I0409 06:16:39.907289 12249 sgd_solver.cpp:106] Iteration 118500, lr = 0.00830714 | |
| I0409 06:22:13.289268 12249 solver.cpp:229] Iteration 119000, loss = 2.61402 | |
| I0409 06:22:13.289566 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.326923 | |
| I0409 06:22:13.289587 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 06:22:13.289600 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.596154 | |
| I0409 06:22:13.289618 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.6664 (* 0.3 = 0.799919 loss) | |
| I0409 06:22:13.289633 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.857286 (* 0.3 = 0.257186 loss) | |
| I0409 06:22:13.289646 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.403846 | |
| I0409 06:22:13.289659 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 06:22:13.289671 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.711538 | |
| I0409 06:22:13.289685 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.49562 (* 0.3 = 0.748685 loss) | |
| I0409 06:22:13.289700 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.762245 (* 0.3 = 0.228674 loss) | |
| I0409 06:22:13.289712 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.692308 | |
| I0409 06:22:13.289724 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.903409 | |
| I0409 06:22:13.289736 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.807692 | |
| I0409 06:22:13.289752 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.78733 (* 1 = 1.78733 loss) | |
| I0409 06:22:13.289767 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.541037 (* 1 = 0.541037 loss) | |
| I0409 06:22:13.289779 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 06:22:13.289791 12249 solver.cpp:245] Train net output #16: total_confidence = 0.401963 | |
| I0409 06:22:13.289806 12249 sgd_solver.cpp:106] Iteration 119000, lr = 0.0083 | |
| I0409 06:27:46.649063 12249 solver.cpp:229] Iteration 119500, loss = 2.67059 | |
| I0409 06:27:46.649215 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.361702 | |
| I0409 06:27:46.649237 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 06:27:46.649250 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.595745 | |
| I0409 06:27:46.649266 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.13078 (* 0.3 = 0.639235 loss) | |
| I0409 06:27:46.649281 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.637823 (* 0.3 = 0.191347 loss) | |
| I0409 06:27:46.649294 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.574468 | |
| I0409 06:27:46.649307 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 06:27:46.649318 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.787234 | |
| I0409 06:27:46.649333 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.44916 (* 0.3 = 0.434748 loss) | |
| I0409 06:27:46.649348 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.430662 (* 0.3 = 0.129199 loss) | |
| I0409 06:27:46.649359 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.87234 | |
| I0409 06:27:46.649371 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 06:27:46.649384 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0409 06:27:46.649397 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.492358 (* 1 = 0.492358 loss) | |
| I0409 06:27:46.649411 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.172659 (* 1 = 0.172659 loss) | |
| I0409 06:27:46.649425 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 06:27:46.649436 12249 solver.cpp:245] Train net output #16: total_confidence = 0.317622 | |
| I0409 06:27:46.649451 12249 sgd_solver.cpp:106] Iteration 119500, lr = 0.00829286 | |
| I0409 06:33:19.960094 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_120000.caffemodel | |
| I0409 06:33:20.398789 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_120000.solverstate | |
| I0409 06:33:20.637243 12249 solver.cpp:338] Iteration 120000, Testing net (#0) | |
| I0409 06:34:01.613256 12249 solver.cpp:393] Test loss: 2.26209 | |
| I0409 06:34:01.613373 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.403778 | |
| I0409 06:34:01.613391 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.846729 | |
| I0409 06:34:01.613404 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.718307 | |
| I0409 06:34:01.613420 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.97453 (* 0.3 = 0.592359 loss) | |
| I0409 06:34:01.613435 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.523324 (* 0.3 = 0.156997 loss) | |
| I0409 06:34:01.613447 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.623445 | |
| I0409 06:34:01.613459 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.897366 | |
| I0409 06:34:01.613471 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.868129 | |
| I0409 06:34:01.613486 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.29665 (* 0.3 = 0.388997 loss) | |
| I0409 06:34:01.613499 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.355162 (* 0.3 = 0.106548 loss) | |
| I0409 06:34:01.613512 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.807356 | |
| I0409 06:34:01.613523 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.950001 | |
| I0409 06:34:01.613534 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.911078 | |
| I0409 06:34:01.613548 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.8004 (* 1 = 0.8004 loss) | |
| I0409 06:34:01.613562 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.216793 (* 1 = 0.216793 loss) | |
| I0409 06:34:01.613574 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.498 | |
| I0409 06:34:01.613585 12249 solver.cpp:406] Test net output #16: total_confidence = 0.423018 | |
| I0409 06:34:01.986304 12249 solver.cpp:229] Iteration 120000, loss = 2.55497 | |
| I0409 06:34:01.986359 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.266667 | |
| I0409 06:34:01.986377 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 06:34:01.986390 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.466667 | |
| I0409 06:34:01.986407 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.86911 (* 0.3 = 0.860732 loss) | |
| I0409 06:34:01.986423 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.78682 (* 0.3 = 0.236046 loss) | |
| I0409 06:34:01.986434 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.466667 | |
| I0409 06:34:01.986448 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 06:34:01.986460 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.688889 | |
| I0409 06:34:01.986474 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.51509 (* 0.3 = 0.754527 loss) | |
| I0409 06:34:01.986488 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.713883 (* 0.3 = 0.214165 loss) | |
| I0409 06:34:01.986501 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.6 | |
| I0409 06:34:01.986513 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0409 06:34:01.986526 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.644444 | |
| I0409 06:34:01.986539 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.45132 (* 1 = 2.45132 loss) | |
| I0409 06:34:01.986554 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.679777 (* 1 = 0.679777 loss) | |
| I0409 06:34:01.986567 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 06:34:01.986579 12249 solver.cpp:245] Train net output #16: total_confidence = 0.259242 | |
| I0409 06:34:01.986593 12249 sgd_solver.cpp:106] Iteration 120000, lr = 0.00828571 | |
| I0409 06:39:35.392350 12249 solver.cpp:229] Iteration 120500, loss = 2.57106 | |
| I0409 06:39:35.392514 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.369565 | |
| I0409 06:39:35.392535 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 06:39:35.392549 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.565217 | |
| I0409 06:39:35.392565 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.22158 (* 0.3 = 0.666475 loss) | |
| I0409 06:39:35.392580 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.636073 (* 0.3 = 0.190822 loss) | |
| I0409 06:39:35.392593 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.521739 | |
| I0409 06:39:35.392606 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 06:39:35.392617 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.804348 | |
| I0409 06:39:35.392632 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.61043 (* 0.3 = 0.483128 loss) | |
| I0409 06:39:35.392647 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.480421 (* 0.3 = 0.144126 loss) | |
| I0409 06:39:35.392658 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.869565 | |
| I0409 06:39:35.392670 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 06:39:35.392683 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0409 06:39:35.392696 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.647246 (* 1 = 0.647246 loss) | |
| I0409 06:39:35.392711 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.175555 (* 1 = 0.175555 loss) | |
| I0409 06:39:35.392724 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 06:39:35.392735 12249 solver.cpp:245] Train net output #16: total_confidence = 0.370299 | |
| I0409 06:39:35.392752 12249 sgd_solver.cpp:106] Iteration 120500, lr = 0.00827857 | |
| I0409 06:45:08.758685 12249 solver.cpp:229] Iteration 121000, loss = 2.63547 | |
| I0409 06:45:08.759003 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.44898 | |
| I0409 06:45:08.759024 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 06:45:08.759037 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.612245 | |
| I0409 06:45:08.759053 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.07367 (* 0.3 = 0.622102 loss) | |
| I0409 06:45:08.759068 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.645149 (* 0.3 = 0.193545 loss) | |
| I0409 06:45:08.759081 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.612245 | |
| I0409 06:45:08.759094 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 06:45:08.759105 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.77551 | |
| I0409 06:45:08.759119 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.55329 (* 0.3 = 0.465987 loss) | |
| I0409 06:45:08.759135 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.470121 (* 0.3 = 0.141036 loss) | |
| I0409 06:45:08.759147 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.857143 | |
| I0409 06:45:08.759160 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 06:45:08.759171 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.938776 | |
| I0409 06:45:08.759186 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.525002 (* 1 = 0.525002 loss) | |
| I0409 06:45:08.759199 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.155821 (* 1 = 0.155821 loss) | |
| I0409 06:45:08.759212 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 06:45:08.759223 12249 solver.cpp:245] Train net output #16: total_confidence = 0.478823 | |
| I0409 06:45:08.759239 12249 sgd_solver.cpp:106] Iteration 121000, lr = 0.00827143 | |
| I0409 06:50:42.140656 12249 solver.cpp:229] Iteration 121500, loss = 2.59114 | |
| I0409 06:50:42.140785 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.288889 | |
| I0409 06:50:42.140806 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 06:50:42.140820 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.622222 | |
| I0409 06:50:42.140836 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.02397 (* 0.3 = 0.607193 loss) | |
| I0409 06:50:42.140853 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.617192 (* 0.3 = 0.185158 loss) | |
| I0409 06:50:42.140866 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.666667 | |
| I0409 06:50:42.140878 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 06:50:42.140890 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.955556 | |
| I0409 06:50:42.140904 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.03829 (* 0.3 = 0.311488 loss) | |
| I0409 06:50:42.140918 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.357376 (* 0.3 = 0.107213 loss) | |
| I0409 06:50:42.140931 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.955556 | |
| I0409 06:50:42.140944 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 06:50:42.140956 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977778 | |
| I0409 06:50:42.140970 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.218969 (* 1 = 0.218969 loss) | |
| I0409 06:50:42.140985 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0618752 (* 1 = 0.0618752 loss) | |
| I0409 06:50:42.140998 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 06:50:42.141010 12249 solver.cpp:245] Train net output #16: total_confidence = 0.561389 | |
| I0409 06:50:42.141024 12249 sgd_solver.cpp:106] Iteration 121500, lr = 0.00826429 | |
| I0409 06:56:15.507429 12249 solver.cpp:229] Iteration 122000, loss = 2.63104 | |
| I0409 06:56:15.507740 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.488372 | |
| I0409 06:56:15.507761 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 06:56:15.507774 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.813953 | |
| I0409 06:56:15.507791 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.66347 (* 0.3 = 0.499042 loss) | |
| I0409 06:56:15.507807 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.508003 (* 0.3 = 0.152401 loss) | |
| I0409 06:56:15.507819 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.651163 | |
| I0409 06:56:15.507833 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 06:56:15.507844 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.883721 | |
| I0409 06:56:15.507858 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.19718 (* 0.3 = 0.359154 loss) | |
| I0409 06:56:15.507872 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.388032 (* 0.3 = 0.11641 loss) | |
| I0409 06:56:15.507885 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.976744 | |
| I0409 06:56:15.507899 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.994318 | |
| I0409 06:56:15.507910 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.976744 | |
| I0409 06:56:15.507925 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.221706 (* 1 = 0.221706 loss) | |
| I0409 06:56:15.507941 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0564779 (* 1 = 0.0564779 loss) | |
| I0409 06:56:15.507953 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 06:56:15.507966 12249 solver.cpp:245] Train net output #16: total_confidence = 0.629188 | |
| I0409 06:56:15.507980 12249 sgd_solver.cpp:106] Iteration 122000, lr = 0.00825714 | |
| I0409 07:01:48.876585 12249 solver.cpp:229] Iteration 122500, loss = 2.61912 | |
| I0409 07:01:48.876822 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.292683 | |
| I0409 07:01:48.876842 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 07:01:48.876854 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.512195 | |
| I0409 07:01:48.876871 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.7901 (* 0.3 = 0.837031 loss) | |
| I0409 07:01:48.876886 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.819586 (* 0.3 = 0.245876 loss) | |
| I0409 07:01:48.876899 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.512195 | |
| I0409 07:01:48.876912 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 07:01:48.876924 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.707317 | |
| I0409 07:01:48.876937 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.03161 (* 0.3 = 0.609483 loss) | |
| I0409 07:01:48.876952 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.603479 (* 0.3 = 0.181044 loss) | |
| I0409 07:01:48.876965 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.780488 | |
| I0409 07:01:48.876977 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 07:01:48.876989 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.804878 | |
| I0409 07:01:48.877003 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.18892 (* 1 = 1.18892 loss) | |
| I0409 07:01:48.877017 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.338719 (* 1 = 0.338719 loss) | |
| I0409 07:01:48.877030 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 07:01:48.877043 12249 solver.cpp:245] Train net output #16: total_confidence = 0.279417 | |
| I0409 07:01:48.877058 12249 sgd_solver.cpp:106] Iteration 122500, lr = 0.00825 | |
| I0409 07:07:22.248556 12249 solver.cpp:229] Iteration 123000, loss = 2.57714 | |
| I0409 07:07:22.248711 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.318182 | |
| I0409 07:07:22.248733 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 07:07:22.248749 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.636364 | |
| I0409 07:07:22.248766 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.08183 (* 0.3 = 0.624548 loss) | |
| I0409 07:07:22.248782 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.550051 (* 0.3 = 0.165015 loss) | |
| I0409 07:07:22.248795 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.545455 | |
| I0409 07:07:22.248808 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 07:07:22.248821 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.840909 | |
| I0409 07:07:22.248836 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.42951 (* 0.3 = 0.428854 loss) | |
| I0409 07:07:22.248849 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.388238 (* 0.3 = 0.116471 loss) | |
| I0409 07:07:22.248862 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.863636 | |
| I0409 07:07:22.248874 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 07:07:22.248888 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.954545 | |
| I0409 07:07:22.248901 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.465583 (* 1 = 0.465583 loss) | |
| I0409 07:07:22.248916 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.132556 (* 1 = 0.132556 loss) | |
| I0409 07:07:22.248929 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 07:07:22.248940 12249 solver.cpp:245] Train net output #16: total_confidence = 0.37236 | |
| I0409 07:07:22.248955 12249 sgd_solver.cpp:106] Iteration 123000, lr = 0.00824286 | |
| I0409 07:12:55.613330 12249 solver.cpp:229] Iteration 123500, loss = 2.56381 | |
| I0409 07:12:55.613579 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.395833 | |
| I0409 07:12:55.613597 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 07:12:55.613611 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6875 | |
| I0409 07:12:55.613627 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95535 (* 0.3 = 0.586606 loss) | |
| I0409 07:12:55.613643 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.556657 (* 0.3 = 0.166997 loss) | |
| I0409 07:12:55.613656 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.6875 | |
| I0409 07:12:55.613668 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 07:12:55.613680 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.895833 | |
| I0409 07:12:55.613694 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.01121 (* 0.3 = 0.303363 loss) | |
| I0409 07:12:55.613708 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.313613 (* 0.3 = 0.094084 loss) | |
| I0409 07:12:55.613721 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.895833 | |
| I0409 07:12:55.613734 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 07:12:55.613745 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 07:12:55.613759 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.364536 (* 1 = 0.364536 loss) | |
| I0409 07:12:55.613773 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.123344 (* 1 = 0.123344 loss) | |
| I0409 07:12:55.613786 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 07:12:55.613798 12249 solver.cpp:245] Train net output #16: total_confidence = 0.490717 | |
| I0409 07:12:55.613813 12249 sgd_solver.cpp:106] Iteration 123500, lr = 0.00823571 | |
| I0409 07:18:28.985133 12249 solver.cpp:229] Iteration 124000, loss = 2.5081 | |
| I0409 07:18:28.985282 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.35 | |
| I0409 07:18:28.985313 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 07:18:28.985337 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.75 | |
| I0409 07:18:28.985355 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.93656 (* 0.3 = 0.580967 loss) | |
| I0409 07:18:28.985370 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.510524 (* 0.3 = 0.153157 loss) | |
| I0409 07:18:28.985383 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55 | |
| I0409 07:18:28.985395 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 07:18:28.985409 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.825 | |
| I0409 07:18:28.985422 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.19924 (* 0.3 = 0.359773 loss) | |
| I0409 07:18:28.985436 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.319846 (* 0.3 = 0.0959539 loss) | |
| I0409 07:18:28.985450 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.95 | |
| I0409 07:18:28.985461 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 07:18:28.985474 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.975 | |
| I0409 07:18:28.985488 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.26638 (* 1 = 0.26638 loss) | |
| I0409 07:18:28.985502 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0648798 (* 1 = 0.0648798 loss) | |
| I0409 07:18:28.985515 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 07:18:28.985527 12249 solver.cpp:245] Train net output #16: total_confidence = 0.484448 | |
| I0409 07:18:28.985543 12249 sgd_solver.cpp:106] Iteration 124000, lr = 0.00822857 | |
| I0409 07:24:02.353322 12249 solver.cpp:229] Iteration 124500, loss = 2.58881 | |
| I0409 07:24:02.353549 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.468085 | |
| I0409 07:24:02.353569 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 07:24:02.353582 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.702128 | |
| I0409 07:24:02.353598 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.00454 (* 0.3 = 0.601361 loss) | |
| I0409 07:24:02.353613 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.589261 (* 0.3 = 0.176778 loss) | |
| I0409 07:24:02.353626 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.531915 | |
| I0409 07:24:02.353638 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 07:24:02.353651 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.765957 | |
| I0409 07:24:02.353664 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.59512 (* 0.3 = 0.478536 loss) | |
| I0409 07:24:02.353678 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.509594 (* 0.3 = 0.152878 loss) | |
| I0409 07:24:02.353691 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.851064 | |
| I0409 07:24:02.353703 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 07:24:02.353715 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0409 07:24:02.353729 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.610039 (* 1 = 0.610039 loss) | |
| I0409 07:24:02.353744 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.208621 (* 1 = 0.208621 loss) | |
| I0409 07:24:02.353755 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 07:24:02.353767 12249 solver.cpp:245] Train net output #16: total_confidence = 0.398221 | |
| I0409 07:24:02.353781 12249 sgd_solver.cpp:106] Iteration 124500, lr = 0.00822143 | |
| I0409 07:29:35.337332 12249 solver.cpp:338] Iteration 125000, Testing net (#0) | |
| I0409 07:30:16.302208 12249 solver.cpp:393] Test loss: 2.22867 | |
| I0409 07:30:16.302331 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.457727 | |
| I0409 07:30:16.302350 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.854912 | |
| I0409 07:30:16.302363 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.739827 | |
| I0409 07:30:16.302379 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.85633 (* 0.3 = 0.556898 loss) | |
| I0409 07:30:16.302395 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.511526 (* 0.3 = 0.153458 loss) | |
| I0409 07:30:16.302407 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.657241 | |
| I0409 07:30:16.302419 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.899594 | |
| I0409 07:30:16.302431 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.87277 | |
| I0409 07:30:16.302444 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.22209 (* 0.3 = 0.366626 loss) | |
| I0409 07:30:16.302459 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.350132 (* 0.3 = 0.10504 loss) | |
| I0409 07:30:16.302470 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.808277 | |
| I0409 07:30:16.302482 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.944546 | |
| I0409 07:30:16.302495 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.908586 | |
| I0409 07:30:16.302507 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.816605 (* 1 = 0.816605 loss) | |
| I0409 07:30:16.302521 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.230042 (* 1 = 0.230042 loss) | |
| I0409 07:30:16.302533 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.467 | |
| I0409 07:30:16.302546 12249 solver.cpp:406] Test net output #16: total_confidence = 0.371828 | |
| I0409 07:30:16.674953 12249 solver.cpp:229] Iteration 125000, loss = 2.53409 | |
| I0409 07:30:16.675009 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.288889 | |
| I0409 07:30:16.675025 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 07:30:16.675039 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.488889 | |
| I0409 07:30:16.675055 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.3939 (* 0.3 = 0.71817 loss) | |
| I0409 07:30:16.675071 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.678025 (* 0.3 = 0.203408 loss) | |
| I0409 07:30:16.675083 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.466667 | |
| I0409 07:30:16.675096 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 07:30:16.675110 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.755556 | |
| I0409 07:30:16.675125 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.64867 (* 0.3 = 0.494601 loss) | |
| I0409 07:30:16.675140 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.475417 (* 0.3 = 0.142625 loss) | |
| I0409 07:30:16.675153 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.8 | |
| I0409 07:30:16.675165 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 07:30:16.675179 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977778 | |
| I0409 07:30:16.675192 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.732124 (* 1 = 0.732124 loss) | |
| I0409 07:30:16.675206 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.196907 (* 1 = 0.196907 loss) | |
| I0409 07:30:16.675218 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 07:30:16.675230 12249 solver.cpp:245] Train net output #16: total_confidence = 0.357864 | |
| I0409 07:30:16.675247 12249 sgd_solver.cpp:106] Iteration 125000, lr = 0.00821429 | |
| I0409 07:35:50.090364 12249 solver.cpp:229] Iteration 125500, loss = 2.47461 | |
| I0409 07:35:50.090673 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.276596 | |
| I0409 07:35:50.090694 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 07:35:50.090708 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.553191 | |
| I0409 07:35:50.090724 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.34911 (* 0.3 = 0.704734 loss) | |
| I0409 07:35:50.090740 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.666509 (* 0.3 = 0.199953 loss) | |
| I0409 07:35:50.090756 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.468085 | |
| I0409 07:35:50.090770 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 07:35:50.090781 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.723404 | |
| I0409 07:35:50.090795 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.57444 (* 0.3 = 0.472332 loss) | |
| I0409 07:35:50.090809 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.44137 (* 0.3 = 0.132411 loss) | |
| I0409 07:35:50.090822 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.808511 | |
| I0409 07:35:50.090834 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 07:35:50.090845 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.893617 | |
| I0409 07:35:50.090859 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.812575 (* 1 = 0.812575 loss) | |
| I0409 07:35:50.090874 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.239658 (* 1 = 0.239658 loss) | |
| I0409 07:35:50.090885 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 07:35:50.090898 12249 solver.cpp:245] Train net output #16: total_confidence = 0.308164 | |
| I0409 07:35:50.090912 12249 sgd_solver.cpp:106] Iteration 125500, lr = 0.00820714 | |
| I0409 07:41:23.467108 12249 solver.cpp:229] Iteration 126000, loss = 2.5125 | |
| I0409 07:41:23.467373 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.411765 | |
| I0409 07:41:23.467394 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 07:41:23.467408 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.686275 | |
| I0409 07:41:23.467424 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.74432 (* 0.3 = 0.523297 loss) | |
| I0409 07:41:23.467439 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.547013 (* 0.3 = 0.164104 loss) | |
| I0409 07:41:23.467453 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.568627 | |
| I0409 07:41:23.467464 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 07:41:23.467476 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.862745 | |
| I0409 07:41:23.467490 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.14446 (* 0.3 = 0.343337 loss) | |
| I0409 07:41:23.467505 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.364198 (* 0.3 = 0.109259 loss) | |
| I0409 07:41:23.467517 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.960784 | |
| I0409 07:41:23.467532 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 07:41:23.467545 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 07:41:23.467560 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.138086 (* 1 = 0.138086 loss) | |
| I0409 07:41:23.467574 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0430773 (* 1 = 0.0430773 loss) | |
| I0409 07:41:23.467586 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 07:41:23.467598 12249 solver.cpp:245] Train net output #16: total_confidence = 0.544787 | |
| I0409 07:41:23.467613 12249 sgd_solver.cpp:106] Iteration 126000, lr = 0.0082 | |
| I0409 07:46:56.841337 12249 solver.cpp:229] Iteration 126500, loss = 2.52097 | |
| I0409 07:46:56.841498 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.428571 | |
| I0409 07:46:56.841519 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 07:46:56.841533 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.761905 | |
| I0409 07:46:56.841549 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.61443 (* 0.3 = 0.484328 loss) | |
| I0409 07:46:56.841565 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.50718 (* 0.3 = 0.152154 loss) | |
| I0409 07:46:56.841578 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.714286 | |
| I0409 07:46:56.841590 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 07:46:56.841603 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.857143 | |
| I0409 07:46:56.841617 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.03899 (* 0.3 = 0.311696 loss) | |
| I0409 07:46:56.841631 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.454881 (* 0.3 = 0.136464 loss) | |
| I0409 07:46:56.841644 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.97619 | |
| I0409 07:46:56.841655 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 07:46:56.841667 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 07:46:56.841682 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.180897 (* 1 = 0.180897 loss) | |
| I0409 07:46:56.841696 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0747018 (* 1 = 0.0747018 loss) | |
| I0409 07:46:56.841709 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 07:46:56.841722 12249 solver.cpp:245] Train net output #16: total_confidence = 0.573108 | |
| I0409 07:46:56.841737 12249 sgd_solver.cpp:106] Iteration 126500, lr = 0.00819286 | |
| I0409 07:52:30.209954 12249 solver.cpp:229] Iteration 127000, loss = 2.52973 | |
| I0409 07:52:30.210222 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0409 07:52:30.210247 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 07:52:30.210261 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.423077 | |
| I0409 07:52:30.210278 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.52605 (* 0.3 = 0.757814 loss) | |
| I0409 07:52:30.210294 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.791788 (* 0.3 = 0.237536 loss) | |
| I0409 07:52:30.210306 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.423077 | |
| I0409 07:52:30.210319 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 07:52:30.210330 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.711538 | |
| I0409 07:52:30.210343 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.95896 (* 0.3 = 0.587688 loss) | |
| I0409 07:52:30.210361 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.596388 (* 0.3 = 0.178916 loss) | |
| I0409 07:52:30.210373 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.865385 | |
| I0409 07:52:30.210386 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 07:52:30.210397 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.961538 | |
| I0409 07:52:30.210412 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.573734 (* 1 = 0.573734 loss) | |
| I0409 07:52:30.210427 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.202048 (* 1 = 0.202048 loss) | |
| I0409 07:52:30.210439 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 07:52:30.210451 12249 solver.cpp:245] Train net output #16: total_confidence = 0.31193 | |
| I0409 07:52:30.210466 12249 sgd_solver.cpp:106] Iteration 127000, lr = 0.00818571 | |
| I0409 07:56:57.836638 12249 sgd_solver.cpp:92] Gradient clipping: scaling down gradients (L2 norm 31.0398 > 30) by scale factor 0.966501 | |
| I0409 07:58:03.580299 12249 solver.cpp:229] Iteration 127500, loss = 2.55627 | |
| I0409 07:58:03.580425 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.463415 | |
| I0409 07:58:03.580446 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 07:58:03.580459 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.731707 | |
| I0409 07:58:03.580476 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.70739 (* 0.3 = 0.512217 loss) | |
| I0409 07:58:03.580492 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.518055 (* 0.3 = 0.155417 loss) | |
| I0409 07:58:03.580503 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.780488 | |
| I0409 07:58:03.580516 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.920455 | |
| I0409 07:58:03.580528 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.902439 | |
| I0409 07:58:03.580543 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.02568 (* 0.3 = 0.307705 loss) | |
| I0409 07:58:03.580570 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.315487 (* 0.3 = 0.0946461 loss) | |
| I0409 07:58:03.580585 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.951219 | |
| I0409 07:58:03.580596 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 07:58:03.580608 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.951219 | |
| I0409 07:58:03.580623 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.291564 (* 1 = 0.291564 loss) | |
| I0409 07:58:03.580637 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0821252 (* 1 = 0.0821252 loss) | |
| I0409 07:58:03.580651 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 07:58:03.580662 12249 solver.cpp:245] Train net output #16: total_confidence = 0.434034 | |
| I0409 07:58:03.580677 12249 sgd_solver.cpp:106] Iteration 127500, lr = 0.00817857 | |
| I0409 08:03:36.945221 12249 solver.cpp:229] Iteration 128000, loss = 2.51403 | |
| I0409 08:03:36.945483 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.347826 | |
| I0409 08:03:36.945504 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 08:03:36.945518 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.543478 | |
| I0409 08:03:36.945535 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.30431 (* 0.3 = 0.691294 loss) | |
| I0409 08:03:36.945550 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.718939 (* 0.3 = 0.215682 loss) | |
| I0409 08:03:36.945562 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.521739 | |
| I0409 08:03:36.945575 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 08:03:36.945587 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.782609 | |
| I0409 08:03:36.945601 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.7256 (* 0.3 = 0.517679 loss) | |
| I0409 08:03:36.945616 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.514302 (* 0.3 = 0.154291 loss) | |
| I0409 08:03:36.945627 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.652174 | |
| I0409 08:03:36.945639 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 08:03:36.945652 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.913043 | |
| I0409 08:03:36.945665 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.12517 (* 1 = 1.12517 loss) | |
| I0409 08:03:36.945679 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.353808 (* 1 = 0.353808 loss) | |
| I0409 08:03:36.945691 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 08:03:36.945703 12249 solver.cpp:245] Train net output #16: total_confidence = 0.335195 | |
| I0409 08:03:36.945718 12249 sgd_solver.cpp:106] Iteration 128000, lr = 0.00817143 | |
| I0409 08:09:10.331878 12249 solver.cpp:229] Iteration 128500, loss = 2.54717 | |
| I0409 08:09:10.332057 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.431818 | |
| I0409 08:09:10.332078 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 08:09:10.332092 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.772727 | |
| I0409 08:09:10.332108 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.64037 (* 0.3 = 0.49211 loss) | |
| I0409 08:09:10.332123 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.495648 (* 0.3 = 0.148694 loss) | |
| I0409 08:09:10.332135 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.681818 | |
| I0409 08:09:10.332149 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 08:09:10.332160 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.954545 | |
| I0409 08:09:10.332175 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.01124 (* 0.3 = 0.303371 loss) | |
| I0409 08:09:10.332190 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.308139 (* 0.3 = 0.0924416 loss) | |
| I0409 08:09:10.332201 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.931818 | |
| I0409 08:09:10.332213 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 08:09:10.332226 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.931818 | |
| I0409 08:09:10.332239 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.418238 (* 1 = 0.418238 loss) | |
| I0409 08:09:10.332254 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.108639 (* 1 = 0.108639 loss) | |
| I0409 08:09:10.332267 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 08:09:10.332280 12249 solver.cpp:245] Train net output #16: total_confidence = 0.567227 | |
| I0409 08:09:10.332295 12249 sgd_solver.cpp:106] Iteration 128500, lr = 0.00816429 | |
| I0409 08:14:44.026849 12249 solver.cpp:229] Iteration 129000, loss = 2.47992 | |
| I0409 08:14:44.027112 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 08:14:44.027133 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 08:14:44.027146 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.62 | |
| I0409 08:14:44.027163 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.20604 (* 0.3 = 0.661812 loss) | |
| I0409 08:14:44.027178 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.646342 (* 0.3 = 0.193902 loss) | |
| I0409 08:14:44.027190 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.42 | |
| I0409 08:14:44.027209 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 08:14:44.027221 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.76 | |
| I0409 08:14:44.027235 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.89614 (* 0.3 = 0.568843 loss) | |
| I0409 08:14:44.027251 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.558536 (* 0.3 = 0.167561 loss) | |
| I0409 08:14:44.027262 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.88 | |
| I0409 08:14:44.027276 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 08:14:44.027287 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.92 | |
| I0409 08:14:44.027302 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.707176 (* 1 = 0.707176 loss) | |
| I0409 08:14:44.027315 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.206338 (* 1 = 0.206338 loss) | |
| I0409 08:14:44.027328 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 08:14:44.027339 12249 solver.cpp:245] Train net output #16: total_confidence = 0.394778 | |
| I0409 08:14:44.027354 12249 sgd_solver.cpp:106] Iteration 129000, lr = 0.00815714 | |
| I0409 08:20:17.387351 12249 solver.cpp:229] Iteration 129500, loss = 2.49319 | |
| I0409 08:20:17.387502 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.302326 | |
| I0409 08:20:17.387522 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 08:20:17.387537 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.55814 | |
| I0409 08:20:17.387552 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.41054 (* 0.3 = 0.723163 loss) | |
| I0409 08:20:17.387567 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.684627 (* 0.3 = 0.205388 loss) | |
| I0409 08:20:17.387580 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.395349 | |
| I0409 08:20:17.387593 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 08:20:17.387605 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.767442 | |
| I0409 08:20:17.387619 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.85825 (* 0.3 = 0.557475 loss) | |
| I0409 08:20:17.387634 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.577084 (* 0.3 = 0.173125 loss) | |
| I0409 08:20:17.387645 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.72093 | |
| I0409 08:20:17.387657 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 08:20:17.387670 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.860465 | |
| I0409 08:20:17.387683 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.03058 (* 1 = 1.03058 loss) | |
| I0409 08:20:17.387697 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.309942 (* 1 = 0.309942 loss) | |
| I0409 08:20:17.387711 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 08:20:17.387722 12249 solver.cpp:245] Train net output #16: total_confidence = 0.208537 | |
| I0409 08:20:17.387737 12249 sgd_solver.cpp:106] Iteration 129500, lr = 0.00815 | |
| I0409 08:25:50.373987 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_130000.caffemodel | |
| I0409 08:25:50.810255 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_130000.solverstate | |
| I0409 08:25:51.048702 12249 solver.cpp:338] Iteration 130000, Testing net (#0) | |
| I0409 08:26:32.009665 12249 solver.cpp:393] Test loss: 2.2589 | |
| I0409 08:26:32.009791 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.432437 | |
| I0409 08:26:32.009810 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.85373 | |
| I0409 08:26:32.009824 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.714287 | |
| I0409 08:26:32.009840 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.99129 (* 0.3 = 0.597389 loss) | |
| I0409 08:26:32.009855 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.52001 (* 0.3 = 0.156003 loss) | |
| I0409 08:26:32.009867 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.620485 | |
| I0409 08:26:32.009879 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.901821 | |
| I0409 08:26:32.009891 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.863702 | |
| I0409 08:26:32.009904 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.32181 (* 0.3 = 0.396542 loss) | |
| I0409 08:26:32.009918 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.3456 (* 0.3 = 0.10368 loss) | |
| I0409 08:26:32.009930 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.811306 | |
| I0409 08:26:32.009943 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.951773 | |
| I0409 08:26:32.009954 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.908748 | |
| I0409 08:26:32.009968 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.791398 (* 1 = 0.791398 loss) | |
| I0409 08:26:32.009981 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.213885 (* 1 = 0.213885 loss) | |
| I0409 08:26:32.009994 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.511 | |
| I0409 08:26:32.010005 12249 solver.cpp:406] Test net output #16: total_confidence = 0.375729 | |
| I0409 08:26:32.383299 12249 solver.cpp:229] Iteration 130000, loss = 2.51094 | |
| I0409 08:26:32.383368 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.380952 | |
| I0409 08:26:32.383386 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 08:26:32.383401 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.547619 | |
| I0409 08:26:32.383417 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.39184 (* 0.3 = 0.717552 loss) | |
| I0409 08:26:32.383432 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.638136 (* 0.3 = 0.191441 loss) | |
| I0409 08:26:32.383445 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.47619 | |
| I0409 08:26:32.383458 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 08:26:32.383471 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.785714 | |
| I0409 08:26:32.383484 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.44969 (* 0.3 = 0.434908 loss) | |
| I0409 08:26:32.383499 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.411231 (* 0.3 = 0.123369 loss) | |
| I0409 08:26:32.383512 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.857143 | |
| I0409 08:26:32.383524 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 08:26:32.383536 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.952381 | |
| I0409 08:26:32.383551 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.455165 (* 1 = 0.455165 loss) | |
| I0409 08:26:32.383566 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.13839 (* 1 = 0.13839 loss) | |
| I0409 08:26:32.383579 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 08:26:32.383590 12249 solver.cpp:245] Train net output #16: total_confidence = 0.356144 | |
| I0409 08:26:32.383605 12249 sgd_solver.cpp:106] Iteration 130000, lr = 0.00814286 | |
| I0409 08:32:05.798238 12249 solver.cpp:229] Iteration 130500, loss = 2.46888 | |
| I0409 08:32:05.798557 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.444444 | |
| I0409 08:32:05.798578 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 08:32:05.798593 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.75 | |
| I0409 08:32:05.798609 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.74602 (* 0.3 = 0.523807 loss) | |
| I0409 08:32:05.798624 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.487738 (* 0.3 = 0.146321 loss) | |
| I0409 08:32:05.798637 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.722222 | |
| I0409 08:32:05.798650 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 08:32:05.798661 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.916667 | |
| I0409 08:32:05.798676 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.01843 (* 0.3 = 0.305529 loss) | |
| I0409 08:32:05.798689 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.367931 (* 0.3 = 0.110379 loss) | |
| I0409 08:32:05.798702 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.916667 | |
| I0409 08:32:05.798714 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 08:32:05.798727 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.972222 | |
| I0409 08:32:05.798741 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.280874 (* 1 = 0.280874 loss) | |
| I0409 08:32:05.798759 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0729467 (* 1 = 0.0729467 loss) | |
| I0409 08:32:05.798773 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 08:32:05.798784 12249 solver.cpp:245] Train net output #16: total_confidence = 0.486939 | |
| I0409 08:32:05.798799 12249 sgd_solver.cpp:106] Iteration 130500, lr = 0.00813571 | |
| I0409 08:37:39.179215 12249 solver.cpp:229] Iteration 131000, loss = 2.45349 | |
| I0409 08:37:39.179335 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.375 | |
| I0409 08:37:39.179355 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 08:37:39.179368 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.7 | |
| I0409 08:37:39.179385 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.90548 (* 0.3 = 0.571643 loss) | |
| I0409 08:37:39.179400 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.512124 (* 0.3 = 0.153637 loss) | |
| I0409 08:37:39.179414 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55 | |
| I0409 08:37:39.179425 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 08:37:39.179437 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.85 | |
| I0409 08:37:39.179451 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.34152 (* 0.3 = 0.402456 loss) | |
| I0409 08:37:39.179466 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.385819 (* 0.3 = 0.115746 loss) | |
| I0409 08:37:39.179478 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.825 | |
| I0409 08:37:39.179491 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 08:37:39.179502 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.95 | |
| I0409 08:37:39.179517 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.660275 (* 1 = 0.660275 loss) | |
| I0409 08:37:39.179532 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.173331 (* 1 = 0.173331 loss) | |
| I0409 08:37:39.179543 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 08:37:39.179555 12249 solver.cpp:245] Train net output #16: total_confidence = 0.392376 | |
| I0409 08:37:39.179569 12249 sgd_solver.cpp:106] Iteration 131000, lr = 0.00812857 | |
| I0409 08:43:12.547541 12249 solver.cpp:229] Iteration 131500, loss = 2.42263 | |
| I0409 08:43:12.547767 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.469388 | |
| I0409 08:43:12.547786 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 08:43:12.547799 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.714286 | |
| I0409 08:43:12.547816 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.78443 (* 0.3 = 0.535329 loss) | |
| I0409 08:43:12.547832 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.536337 (* 0.3 = 0.160901 loss) | |
| I0409 08:43:12.547844 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.612245 | |
| I0409 08:43:12.547857 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 08:43:12.547868 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.816327 | |
| I0409 08:43:12.547883 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.40276 (* 0.3 = 0.420827 loss) | |
| I0409 08:43:12.547897 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.443883 (* 0.3 = 0.133165 loss) | |
| I0409 08:43:12.547910 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.938776 | |
| I0409 08:43:12.547922 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 08:43:12.547935 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 08:43:12.547948 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.168841 (* 1 = 0.168841 loss) | |
| I0409 08:43:12.547963 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0654866 (* 1 = 0.0654866 loss) | |
| I0409 08:43:12.547976 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 08:43:12.547989 12249 solver.cpp:245] Train net output #16: total_confidence = 0.538119 | |
| I0409 08:43:12.548003 12249 sgd_solver.cpp:106] Iteration 131500, lr = 0.00812143 | |
| I0409 08:48:45.915904 12249 solver.cpp:229] Iteration 132000, loss = 2.51872 | |
| I0409 08:48:45.916038 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.255319 | |
| I0409 08:48:45.916057 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 08:48:45.916070 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.617021 | |
| I0409 08:48:45.916087 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.32542 (* 0.3 = 0.697627 loss) | |
| I0409 08:48:45.916102 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.700437 (* 0.3 = 0.210131 loss) | |
| I0409 08:48:45.916115 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.382979 | |
| I0409 08:48:45.916127 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 08:48:45.916141 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.702128 | |
| I0409 08:48:45.916154 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.88819 (* 0.3 = 0.566456 loss) | |
| I0409 08:48:45.916168 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.570185 (* 0.3 = 0.171056 loss) | |
| I0409 08:48:45.916182 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.851064 | |
| I0409 08:48:45.916193 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 08:48:45.916206 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 08:48:45.916220 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.435911 (* 1 = 0.435911 loss) | |
| I0409 08:48:45.916235 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.119915 (* 1 = 0.119915 loss) | |
| I0409 08:48:45.916247 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 08:48:45.916260 12249 solver.cpp:245] Train net output #16: total_confidence = 0.406243 | |
| I0409 08:48:45.916275 12249 sgd_solver.cpp:106] Iteration 132000, lr = 0.00811428 | |
| I0409 08:54:19.287875 12249 solver.cpp:229] Iteration 132500, loss = 2.50314 | |
| I0409 08:54:19.288115 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 08:54:19.288135 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 08:54:19.288147 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.673913 | |
| I0409 08:54:19.288163 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.99246 (* 0.3 = 0.597739 loss) | |
| I0409 08:54:19.288179 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.589025 (* 0.3 = 0.176707 loss) | |
| I0409 08:54:19.288192 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.586957 | |
| I0409 08:54:19.288204 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 08:54:19.288216 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.869565 | |
| I0409 08:54:19.288230 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.32891 (* 0.3 = 0.398672 loss) | |
| I0409 08:54:19.288244 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.449571 (* 0.3 = 0.134871 loss) | |
| I0409 08:54:19.288257 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826087 | |
| I0409 08:54:19.288269 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 08:54:19.288282 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0409 08:54:19.288296 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.515987 (* 1 = 0.515987 loss) | |
| I0409 08:54:19.288311 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.139833 (* 1 = 0.139833 loss) | |
| I0409 08:54:19.288323 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 08:54:19.288336 12249 solver.cpp:245] Train net output #16: total_confidence = 0.523376 | |
| I0409 08:54:19.288350 12249 sgd_solver.cpp:106] Iteration 132500, lr = 0.00810714 | |
| I0409 08:59:52.663269 12249 solver.cpp:229] Iteration 133000, loss = 2.42525 | |
| I0409 08:59:52.663422 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.333333 | |
| I0409 08:59:52.663444 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 08:59:52.663457 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.645833 | |
| I0409 08:59:52.663473 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.17426 (* 0.3 = 0.652277 loss) | |
| I0409 08:59:52.663488 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.644224 (* 0.3 = 0.193267 loss) | |
| I0409 08:59:52.663501 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.520833 | |
| I0409 08:59:52.663513 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 08:59:52.663525 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.854167 | |
| I0409 08:59:52.663539 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.40342 (* 0.3 = 0.421026 loss) | |
| I0409 08:59:52.663553 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.451407 (* 0.3 = 0.135422 loss) | |
| I0409 08:59:52.663566 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.895833 | |
| I0409 08:59:52.663578 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 08:59:52.663590 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.979167 | |
| I0409 08:59:52.663604 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.486336 (* 1 = 0.486336 loss) | |
| I0409 08:59:52.663619 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.141794 (* 1 = 0.141794 loss) | |
| I0409 08:59:52.663631 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 08:59:52.663643 12249 solver.cpp:245] Train net output #16: total_confidence = 0.397453 | |
| I0409 08:59:52.663658 12249 sgd_solver.cpp:106] Iteration 133000, lr = 0.0081 | |
| I0409 09:05:26.027685 12249 solver.cpp:229] Iteration 133500, loss = 2.42622 | |
| I0409 09:05:26.027926 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.320755 | |
| I0409 09:05:26.027945 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 09:05:26.027958 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.622642 | |
| I0409 09:05:26.027974 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.20057 (* 0.3 = 0.660171 loss) | |
| I0409 09:05:26.027990 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.695938 (* 0.3 = 0.208781 loss) | |
| I0409 09:05:26.028003 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.471698 | |
| I0409 09:05:26.028015 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 09:05:26.028028 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.754717 | |
| I0409 09:05:26.028041 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.75669 (* 0.3 = 0.527006 loss) | |
| I0409 09:05:26.028055 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.545657 (* 0.3 = 0.163697 loss) | |
| I0409 09:05:26.028069 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.849057 | |
| I0409 09:05:26.028080 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 09:05:26.028092 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.943396 | |
| I0409 09:05:26.028107 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.485684 (* 1 = 0.485684 loss) | |
| I0409 09:05:26.028121 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.185563 (* 1 = 0.185563 loss) | |
| I0409 09:05:26.028133 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 09:05:26.028146 12249 solver.cpp:245] Train net output #16: total_confidence = 0.323552 | |
| I0409 09:05:26.028159 12249 sgd_solver.cpp:106] Iteration 133500, lr = 0.00809286 | |
| I0409 09:10:59.399304 12249 solver.cpp:229] Iteration 134000, loss = 2.45457 | |
| I0409 09:10:59.399673 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.413043 | |
| I0409 09:10:59.399695 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 09:10:59.399709 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.673913 | |
| I0409 09:10:59.399726 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.00317 (* 0.3 = 0.60095 loss) | |
| I0409 09:10:59.399741 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.59016 (* 0.3 = 0.177048 loss) | |
| I0409 09:10:59.399758 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 09:10:59.399771 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 09:10:59.399783 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.782609 | |
| I0409 09:10:59.399797 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.47147 (* 0.3 = 0.441441 loss) | |
| I0409 09:10:59.399812 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.476818 (* 0.3 = 0.143045 loss) | |
| I0409 09:10:59.399826 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.891304 | |
| I0409 09:10:59.399837 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 09:10:59.399849 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0409 09:10:59.399863 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.409579 (* 1 = 0.409579 loss) | |
| I0409 09:10:59.399878 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.137488 (* 1 = 0.137488 loss) | |
| I0409 09:10:59.399890 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 09:10:59.399902 12249 solver.cpp:245] Train net output #16: total_confidence = 0.455602 | |
| I0409 09:10:59.399917 12249 sgd_solver.cpp:106] Iteration 134000, lr = 0.00808571 | |
| I0409 09:16:32.768645 12249 solver.cpp:229] Iteration 134500, loss = 2.54263 | |
| I0409 09:16:32.768771 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.261905 | |
| I0409 09:16:32.768791 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 09:16:32.768805 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.428571 | |
| I0409 09:16:32.768821 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.22645 (* 0.3 = 0.667935 loss) | |
| I0409 09:16:32.768836 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.588784 (* 0.3 = 0.176635 loss) | |
| I0409 09:16:32.768849 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.52381 | |
| I0409 09:16:32.768862 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 09:16:32.768873 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.714286 | |
| I0409 09:16:32.768887 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.6016 (* 0.3 = 0.48048 loss) | |
| I0409 09:16:32.768901 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.409878 (* 0.3 = 0.122963 loss) | |
| I0409 09:16:32.768914 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.761905 | |
| I0409 09:16:32.768926 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 09:16:32.768939 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 09:16:32.768952 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.60555 (* 1 = 0.60555 loss) | |
| I0409 09:16:32.768966 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.14947 (* 1 = 0.14947 loss) | |
| I0409 09:16:32.768980 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 09:16:32.768991 12249 solver.cpp:245] Train net output #16: total_confidence = 0.281216 | |
| I0409 09:16:32.769006 12249 sgd_solver.cpp:106] Iteration 134500, lr = 0.00807857 | |
| I0409 09:22:05.746928 12249 solver.cpp:338] Iteration 135000, Testing net (#0) | |
| I0409 09:22:46.827841 12249 solver.cpp:393] Test loss: 2.1199 | |
| I0409 09:22:46.827972 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.453829 | |
| I0409 09:22:46.827993 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.861002 | |
| I0409 09:22:46.828007 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.750279 | |
| I0409 09:22:46.828024 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.80973 (* 0.3 = 0.542919 loss) | |
| I0409 09:22:46.828040 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.46949 (* 0.3 = 0.140847 loss) | |
| I0409 09:22:46.828052 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.668795 | |
| I0409 09:22:46.828064 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.90923 | |
| I0409 09:22:46.828076 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.885917 | |
| I0409 09:22:46.828090 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.15724 (* 0.3 = 0.347171 loss) | |
| I0409 09:22:46.828109 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.31592 (* 0.3 = 0.0947759 loss) | |
| I0409 09:22:46.828121 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.81235 | |
| I0409 09:22:46.828133 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.954136 | |
| I0409 09:22:46.828145 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.913773 | |
| I0409 09:22:46.828158 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.794635 (* 1 = 0.794635 loss) | |
| I0409 09:22:46.828172 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.199556 (* 1 = 0.199556 loss) | |
| I0409 09:22:46.828184 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.526 | |
| I0409 09:22:46.828197 12249 solver.cpp:406] Test net output #16: total_confidence = 0.44173 | |
| I0409 09:22:47.204591 12249 solver.cpp:229] Iteration 135000, loss = 2.43466 | |
| I0409 09:22:47.204669 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.295455 | |
| I0409 09:22:47.204687 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 09:22:47.204701 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.545455 | |
| I0409 09:22:47.204718 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.26785 (* 0.3 = 0.680355 loss) | |
| I0409 09:22:47.204733 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.605044 (* 0.3 = 0.181513 loss) | |
| I0409 09:22:47.204747 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.545455 | |
| I0409 09:22:47.204759 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 09:22:47.204771 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.840909 | |
| I0409 09:22:47.204787 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.55497 (* 0.3 = 0.466491 loss) | |
| I0409 09:22:47.204800 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.415797 (* 0.3 = 0.124739 loss) | |
| I0409 09:22:47.204813 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.840909 | |
| I0409 09:22:47.204825 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 09:22:47.204838 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.954545 | |
| I0409 09:22:47.204852 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.52457 (* 1 = 0.52457 loss) | |
| I0409 09:22:47.204867 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.133559 (* 1 = 0.133559 loss) | |
| I0409 09:22:47.204879 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 09:22:47.204892 12249 solver.cpp:245] Train net output #16: total_confidence = 0.297796 | |
| I0409 09:22:47.204907 12249 sgd_solver.cpp:106] Iteration 135000, lr = 0.00807143 | |
| I0409 09:28:21.171994 12249 solver.cpp:229] Iteration 135500, loss = 2.44013 | |
| I0409 09:28:21.172157 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.380952 | |
| I0409 09:28:21.172176 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 09:28:21.172190 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.52381 | |
| I0409 09:28:21.172207 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.43216 (* 0.3 = 0.729649 loss) | |
| I0409 09:28:21.172222 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.648775 (* 0.3 = 0.194633 loss) | |
| I0409 09:28:21.172235 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.52381 | |
| I0409 09:28:21.172247 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 09:28:21.172260 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.809524 | |
| I0409 09:28:21.172274 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.59839 (* 0.3 = 0.479516 loss) | |
| I0409 09:28:21.172289 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.474421 (* 0.3 = 0.142326 loss) | |
| I0409 09:28:21.172302 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.833333 | |
| I0409 09:28:21.172313 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 09:28:21.172325 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.904762 | |
| I0409 09:28:21.172339 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.975585 (* 1 = 0.975585 loss) | |
| I0409 09:28:21.172353 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.237155 (* 1 = 0.237155 loss) | |
| I0409 09:28:21.172366 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 09:28:21.172379 12249 solver.cpp:245] Train net output #16: total_confidence = 0.301255 | |
| I0409 09:28:21.172394 12249 sgd_solver.cpp:106] Iteration 135500, lr = 0.00806428 | |
| I0409 09:33:54.559824 12249 solver.cpp:229] Iteration 136000, loss = 2.43957 | |
| I0409 09:33:54.560108 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.52381 | |
| I0409 09:33:54.560128 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 09:33:54.560142 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.761905 | |
| I0409 09:33:54.560158 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.76305 (* 0.3 = 0.528916 loss) | |
| I0409 09:33:54.560173 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.568276 (* 0.3 = 0.170483 loss) | |
| I0409 09:33:54.560186 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.595238 | |
| I0409 09:33:54.560199 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 09:33:54.560211 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.928571 | |
| I0409 09:33:54.560226 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.14208 (* 0.3 = 0.342625 loss) | |
| I0409 09:33:54.560241 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.372309 (* 0.3 = 0.111693 loss) | |
| I0409 09:33:54.560253 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.904762 | |
| I0409 09:33:54.560266 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 09:33:54.560278 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.97619 | |
| I0409 09:33:54.560292 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.397228 (* 1 = 0.397228 loss) | |
| I0409 09:33:54.560307 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.115881 (* 1 = 0.115881 loss) | |
| I0409 09:33:54.560320 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 09:33:54.560335 12249 solver.cpp:245] Train net output #16: total_confidence = 0.427768 | |
| I0409 09:33:54.560351 12249 sgd_solver.cpp:106] Iteration 136000, lr = 0.00805714 | |
| I0409 09:39:27.921391 12249 solver.cpp:229] Iteration 136500, loss = 2.42259 | |
| I0409 09:39:27.921540 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 09:39:27.921561 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 09:39:27.921573 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.673913 | |
| I0409 09:39:27.921589 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.91763 (* 0.3 = 0.575289 loss) | |
| I0409 09:39:27.921604 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.535481 (* 0.3 = 0.160644 loss) | |
| I0409 09:39:27.921617 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.456522 | |
| I0409 09:39:27.921630 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 09:39:27.921643 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.826087 | |
| I0409 09:39:27.921656 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.62269 (* 0.3 = 0.486807 loss) | |
| I0409 09:39:27.921670 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.448206 (* 0.3 = 0.134462 loss) | |
| I0409 09:39:27.921682 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.76087 | |
| I0409 09:39:27.921695 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 09:39:27.921707 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0409 09:39:27.921721 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.792635 (* 1 = 0.792635 loss) | |
| I0409 09:39:27.921736 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.221793 (* 1 = 0.221793 loss) | |
| I0409 09:39:27.921751 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 09:39:27.921764 12249 solver.cpp:245] Train net output #16: total_confidence = 0.413552 | |
| I0409 09:39:27.921779 12249 sgd_solver.cpp:106] Iteration 136500, lr = 0.00805 | |
| I0409 09:45:01.294555 12249 solver.cpp:229] Iteration 137000, loss = 2.4345 | |
| I0409 09:45:01.294790 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.333333 | |
| I0409 09:45:01.294811 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 09:45:01.294824 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.555556 | |
| I0409 09:45:01.294841 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.39231 (* 0.3 = 0.717692 loss) | |
| I0409 09:45:01.294855 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.684899 (* 0.3 = 0.20547 loss) | |
| I0409 09:45:01.294868 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4 | |
| I0409 09:45:01.294880 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 09:45:01.294893 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0409 09:45:01.294906 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.14499 (* 0.3 = 0.643496 loss) | |
| I0409 09:45:01.294920 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.652116 (* 0.3 = 0.195635 loss) | |
| I0409 09:45:01.294934 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.577778 | |
| I0409 09:45:01.294945 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0409 09:45:01.294957 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.777778 | |
| I0409 09:45:01.294972 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.6171 (* 1 = 1.6171 loss) | |
| I0409 09:45:01.294986 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.466087 (* 1 = 0.466087 loss) | |
| I0409 09:45:01.294998 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 09:45:01.295011 12249 solver.cpp:245] Train net output #16: total_confidence = 0.298845 | |
| I0409 09:45:01.295025 12249 sgd_solver.cpp:106] Iteration 137000, lr = 0.00804286 | |
| I0409 09:50:34.670027 12249 solver.cpp:229] Iteration 137500, loss = 2.44517 | |
| I0409 09:50:34.670187 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 09:50:34.670208 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 09:50:34.670222 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.68 | |
| I0409 09:50:34.670238 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95028 (* 0.3 = 0.585084 loss) | |
| I0409 09:50:34.670253 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.597315 (* 0.3 = 0.179195 loss) | |
| I0409 09:50:34.670265 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.56 | |
| I0409 09:50:34.670279 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 09:50:34.670290 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.88 | |
| I0409 09:50:34.670305 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.36227 (* 0.3 = 0.408681 loss) | |
| I0409 09:50:34.670318 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.414429 (* 0.3 = 0.124329 loss) | |
| I0409 09:50:34.670331 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.92 | |
| I0409 09:50:34.670343 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 09:50:34.670356 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.98 | |
| I0409 09:50:34.670369 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.378492 (* 1 = 0.378492 loss) | |
| I0409 09:50:34.670383 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.122245 (* 1 = 0.122245 loss) | |
| I0409 09:50:34.670397 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 09:50:34.670408 12249 solver.cpp:245] Train net output #16: total_confidence = 0.300618 | |
| I0409 09:50:34.670423 12249 sgd_solver.cpp:106] Iteration 137500, lr = 0.00803571 | |
| I0409 09:56:08.026155 12249 solver.cpp:229] Iteration 138000, loss = 2.43391 | |
| I0409 09:56:08.026391 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.395833 | |
| I0409 09:56:08.026410 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 09:56:08.026423 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6875 | |
| I0409 09:56:08.026440 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.00092 (* 0.3 = 0.600276 loss) | |
| I0409 09:56:08.026455 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.628776 (* 0.3 = 0.188633 loss) | |
| I0409 09:56:08.026468 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 09:56:08.026481 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 09:56:08.026494 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.833333 | |
| I0409 09:56:08.026507 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.53568 (* 0.3 = 0.460703 loss) | |
| I0409 09:56:08.026521 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.449136 (* 0.3 = 0.134741 loss) | |
| I0409 09:56:08.026535 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.875 | |
| I0409 09:56:08.026546 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 09:56:08.026559 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.958333 | |
| I0409 09:56:08.026573 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.566222 (* 1 = 0.566222 loss) | |
| I0409 09:56:08.026588 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.179015 (* 1 = 0.179015 loss) | |
| I0409 09:56:08.026602 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 09:56:08.026613 12249 solver.cpp:245] Train net output #16: total_confidence = 0.253494 | |
| I0409 09:56:08.026628 12249 sgd_solver.cpp:106] Iteration 138000, lr = 0.00802857 | |
| I0409 10:01:41.402386 12249 solver.cpp:229] Iteration 138500, loss = 2.42245 | |
| I0409 10:01:41.402679 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.25 | |
| I0409 10:01:41.402699 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 10:01:41.402714 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.431818 | |
| I0409 10:01:41.402730 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.59756 (* 0.3 = 0.779269 loss) | |
| I0409 10:01:41.402747 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.761828 (* 0.3 = 0.228548 loss) | |
| I0409 10:01:41.402761 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.477273 | |
| I0409 10:01:41.402775 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 10:01:41.402786 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.727273 | |
| I0409 10:01:41.402801 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.81103 (* 0.3 = 0.543309 loss) | |
| I0409 10:01:41.402814 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.567566 (* 0.3 = 0.17027 loss) | |
| I0409 10:01:41.402827 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.636364 | |
| I0409 10:01:41.402839 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 10:01:41.402851 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.75 | |
| I0409 10:01:41.402865 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.2007 (* 1 = 1.2007 loss) | |
| I0409 10:01:41.402879 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.360287 (* 1 = 0.360287 loss) | |
| I0409 10:01:41.402892 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 10:01:41.402904 12249 solver.cpp:245] Train net output #16: total_confidence = 0.296864 | |
| I0409 10:01:41.402918 12249 sgd_solver.cpp:106] Iteration 138500, lr = 0.00802143 | |
| I0409 10:07:14.778046 12249 solver.cpp:229] Iteration 139000, loss = 2.41146 | |
| I0409 10:07:14.778178 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.333333 | |
| I0409 10:07:14.778199 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 10:07:14.778213 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.627451 | |
| I0409 10:07:14.778229 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.1651 (* 0.3 = 0.64953 loss) | |
| I0409 10:07:14.778244 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.637674 (* 0.3 = 0.191302 loss) | |
| I0409 10:07:14.778259 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.431373 | |
| I0409 10:07:14.778270 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 10:07:14.778282 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.803922 | |
| I0409 10:07:14.778296 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.83572 (* 0.3 = 0.550717 loss) | |
| I0409 10:07:14.778311 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.543552 (* 0.3 = 0.163066 loss) | |
| I0409 10:07:14.778324 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.72549 | |
| I0409 10:07:14.778337 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 10:07:14.778348 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.843137 | |
| I0409 10:07:14.778363 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.943864 (* 1 = 0.943864 loss) | |
| I0409 10:07:14.778378 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.281438 (* 1 = 0.281438 loss) | |
| I0409 10:07:14.778390 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 10:07:14.778403 12249 solver.cpp:245] Train net output #16: total_confidence = 0.306502 | |
| I0409 10:07:14.778416 12249 sgd_solver.cpp:106] Iteration 139000, lr = 0.00801429 | |
| I0409 10:12:48.145740 12249 solver.cpp:229] Iteration 139500, loss = 2.47688 | |
| I0409 10:12:48.146035 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.245283 | |
| I0409 10:12:48.146056 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 10:12:48.146070 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.566038 | |
| I0409 10:12:48.146085 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.3621 (* 0.3 = 0.708629 loss) | |
| I0409 10:12:48.146101 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.727688 (* 0.3 = 0.218306 loss) | |
| I0409 10:12:48.146114 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.45283 | |
| I0409 10:12:48.146127 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 10:12:48.146139 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.698113 | |
| I0409 10:12:48.146153 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.69803 (* 0.3 = 0.509409 loss) | |
| I0409 10:12:48.146167 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.534186 (* 0.3 = 0.160256 loss) | |
| I0409 10:12:48.146180 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.792453 | |
| I0409 10:12:48.146193 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 10:12:48.146204 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.867925 | |
| I0409 10:12:48.146219 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.777967 (* 1 = 0.777967 loss) | |
| I0409 10:12:48.146234 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.250802 (* 1 = 0.250802 loss) | |
| I0409 10:12:48.146245 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 10:12:48.146257 12249 solver.cpp:245] Train net output #16: total_confidence = 0.31746 | |
| I0409 10:12:48.146272 12249 sgd_solver.cpp:106] Iteration 139500, lr = 0.00800714 | |
| I0409 10:18:21.133306 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_140000.caffemodel | |
| I0409 10:18:21.595228 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_140000.solverstate | |
| I0409 10:18:21.835618 12249 solver.cpp:338] Iteration 140000, Testing net (#0) | |
| I0409 10:19:02.801059 12249 solver.cpp:393] Test loss: 2.20881 | |
| I0409 10:19:02.801162 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.442424 | |
| I0409 10:19:02.801182 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.855912 | |
| I0409 10:19:02.801194 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.748575 | |
| I0409 10:19:02.801210 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.85039 (* 0.3 = 0.555118 loss) | |
| I0409 10:19:02.801225 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.487163 (* 0.3 = 0.146149 loss) | |
| I0409 10:19:02.801239 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.665583 | |
| I0409 10:19:02.801249 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.910275 | |
| I0409 10:19:02.801261 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.879537 | |
| I0409 10:19:02.801275 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.20257 (* 0.3 = 0.360772 loss) | |
| I0409 10:19:02.801288 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.320179 (* 0.3 = 0.0960538 loss) | |
| I0409 10:19:02.801301 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.813651 | |
| I0409 10:19:02.801312 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.953137 | |
| I0409 10:19:02.801323 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.906417 | |
| I0409 10:19:02.801337 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.836632 (* 1 = 0.836632 loss) | |
| I0409 10:19:02.801352 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.214086 (* 1 = 0.214086 loss) | |
| I0409 10:19:02.801363 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.524 | |
| I0409 10:19:02.801374 12249 solver.cpp:406] Test net output #16: total_confidence = 0.487279 | |
| I0409 10:19:03.174015 12249 solver.cpp:229] Iteration 140000, loss = 2.42223 | |
| I0409 10:19:03.174064 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.368421 | |
| I0409 10:19:03.174082 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 10:19:03.174094 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.605263 | |
| I0409 10:19:03.174109 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.35258 (* 0.3 = 0.705775 loss) | |
| I0409 10:19:03.174124 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.646389 (* 0.3 = 0.193917 loss) | |
| I0409 10:19:03.174137 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.526316 | |
| I0409 10:19:03.174151 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 10:19:03.174173 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.789474 | |
| I0409 10:19:03.174199 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.84627 (* 0.3 = 0.55388 loss) | |
| I0409 10:19:03.174226 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.51474 (* 0.3 = 0.154422 loss) | |
| I0409 10:19:03.174255 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.684211 | |
| I0409 10:19:03.174271 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 10:19:03.174283 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.789474 | |
| I0409 10:19:03.174298 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.3984 (* 1 = 1.3984 loss) | |
| I0409 10:19:03.174312 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.333392 (* 1 = 0.333392 loss) | |
| I0409 10:19:03.174324 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 10:19:03.174336 12249 solver.cpp:245] Train net output #16: total_confidence = 0.340496 | |
| I0409 10:19:03.174351 12249 sgd_solver.cpp:106] Iteration 140000, lr = 0.008 | |
| I0409 10:24:36.542121 12249 solver.cpp:229] Iteration 140500, loss = 2.38628 | |
| I0409 10:24:36.542440 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.408163 | |
| I0409 10:24:36.542471 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 10:24:36.542495 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.632653 | |
| I0409 10:24:36.542526 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.17745 (* 0.3 = 0.653236 loss) | |
| I0409 10:24:36.542553 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.653935 (* 0.3 = 0.196181 loss) | |
| I0409 10:24:36.542575 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.469388 | |
| I0409 10:24:36.542598 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 10:24:36.542621 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.755102 | |
| I0409 10:24:36.542651 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.7268 (* 0.3 = 0.518041 loss) | |
| I0409 10:24:36.542680 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.519459 (* 0.3 = 0.155838 loss) | |
| I0409 10:24:36.542703 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.795918 | |
| I0409 10:24:36.542726 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 10:24:36.542752 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.877551 | |
| I0409 10:24:36.542779 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.817548 (* 1 = 0.817548 loss) | |
| I0409 10:24:36.542805 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.252092 (* 1 = 0.252092 loss) | |
| I0409 10:24:36.542829 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 10:24:36.542851 12249 solver.cpp:245] Train net output #16: total_confidence = 0.315377 | |
| I0409 10:24:36.542876 12249 sgd_solver.cpp:106] Iteration 140500, lr = 0.00799286 | |
| I0409 10:30:09.922626 12249 solver.cpp:229] Iteration 141000, loss = 2.38396 | |
| I0409 10:30:09.922776 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.27907 | |
| I0409 10:30:09.922797 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 10:30:09.922811 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.604651 | |
| I0409 10:30:09.922827 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.3115 (* 0.3 = 0.693449 loss) | |
| I0409 10:30:09.922842 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.700441 (* 0.3 = 0.210132 loss) | |
| I0409 10:30:09.922854 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.488372 | |
| I0409 10:30:09.922868 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 10:30:09.922879 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.790698 | |
| I0409 10:30:09.922894 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.76373 (* 0.3 = 0.52912 loss) | |
| I0409 10:30:09.922907 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.564903 (* 0.3 = 0.169471 loss) | |
| I0409 10:30:09.922919 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.813953 | |
| I0409 10:30:09.922931 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 10:30:09.922943 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 10:30:09.922957 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.592549 (* 1 = 0.592549 loss) | |
| I0409 10:30:09.922971 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.209952 (* 1 = 0.209952 loss) | |
| I0409 10:30:09.922983 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 10:30:09.922996 12249 solver.cpp:245] Train net output #16: total_confidence = 0.389136 | |
| I0409 10:30:09.923019 12249 sgd_solver.cpp:106] Iteration 141000, lr = 0.00798571 | |
| I0409 10:35:43.286412 12249 solver.cpp:229] Iteration 141500, loss = 2.39582 | |
| I0409 10:35:43.286726 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.480769 | |
| I0409 10:35:43.286751 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 10:35:43.286764 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.788462 | |
| I0409 10:35:43.286782 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.85391 (* 0.3 = 0.556172 loss) | |
| I0409 10:35:43.286797 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.607594 (* 0.3 = 0.182278 loss) | |
| I0409 10:35:43.286809 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.576923 | |
| I0409 10:35:43.286823 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 10:35:43.286834 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.807692 | |
| I0409 10:35:43.286849 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.42331 (* 0.3 = 0.426994 loss) | |
| I0409 10:35:43.286862 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.444153 (* 0.3 = 0.133246 loss) | |
| I0409 10:35:43.286875 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826923 | |
| I0409 10:35:43.286888 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 10:35:43.286900 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.923077 | |
| I0409 10:35:43.286916 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.646671 (* 1 = 0.646671 loss) | |
| I0409 10:35:43.286929 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.201488 (* 1 = 0.201488 loss) | |
| I0409 10:35:43.286942 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 10:35:43.286954 12249 solver.cpp:245] Train net output #16: total_confidence = 0.365466 | |
| I0409 10:35:43.286968 12249 sgd_solver.cpp:106] Iteration 141500, lr = 0.00797857 | |
| I0409 10:41:16.661661 12249 solver.cpp:229] Iteration 142000, loss = 2.40584 | |
| I0409 10:41:16.661986 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 10:41:16.662008 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 10:41:16.662021 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.869565 | |
| I0409 10:41:16.662039 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.53665 (* 0.3 = 0.460996 loss) | |
| I0409 10:41:16.662053 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.471956 (* 0.3 = 0.141587 loss) | |
| I0409 10:41:16.662066 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.673913 | |
| I0409 10:41:16.662078 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 10:41:16.662091 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.956522 | |
| I0409 10:41:16.662104 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.921542 (* 0.3 = 0.276463 loss) | |
| I0409 10:41:16.662118 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.264772 (* 0.3 = 0.0794316 loss) | |
| I0409 10:41:16.662132 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.934783 | |
| I0409 10:41:16.662143 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 10:41:16.662155 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 10:41:16.662169 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.328764 (* 1 = 0.328764 loss) | |
| I0409 10:41:16.662184 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.102598 (* 1 = 0.102598 loss) | |
| I0409 10:41:16.662196 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 10:41:16.662209 12249 solver.cpp:245] Train net output #16: total_confidence = 0.578115 | |
| I0409 10:41:16.662223 12249 sgd_solver.cpp:106] Iteration 142000, lr = 0.00797143 | |
| I0409 10:46:50.031373 12249 solver.cpp:229] Iteration 142500, loss = 2.38421 | |
| I0409 10:46:50.031479 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 10:46:50.031498 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 10:46:50.031512 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6 | |
| I0409 10:46:50.031528 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.1332 (* 0.3 = 0.63996 loss) | |
| I0409 10:46:50.031543 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.627674 (* 0.3 = 0.188302 loss) | |
| I0409 10:46:50.031556 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.6 | |
| I0409 10:46:50.031569 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 10:46:50.031580 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.866667 | |
| I0409 10:46:50.031594 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.43086 (* 0.3 = 0.429259 loss) | |
| I0409 10:46:50.031608 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.472932 (* 0.3 = 0.14188 loss) | |
| I0409 10:46:50.031620 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.844444 | |
| I0409 10:46:50.031632 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 10:46:50.031644 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.866667 | |
| I0409 10:46:50.031658 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.769212 (* 1 = 0.769212 loss) | |
| I0409 10:46:50.031672 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.224479 (* 1 = 0.224479 loss) | |
| I0409 10:46:50.031684 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 10:46:50.031697 12249 solver.cpp:245] Train net output #16: total_confidence = 0.517022 | |
| I0409 10:46:50.031711 12249 sgd_solver.cpp:106] Iteration 142500, lr = 0.00796429 | |
| I0409 10:52:23.408284 12249 solver.cpp:229] Iteration 143000, loss = 2.42997 | |
| I0409 10:52:23.408656 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.219512 | |
| I0409 10:52:23.408679 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 10:52:23.408692 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.536585 | |
| I0409 10:52:23.408709 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.1548 (* 0.3 = 0.646439 loss) | |
| I0409 10:52:23.408725 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.571644 (* 0.3 = 0.171493 loss) | |
| I0409 10:52:23.408736 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.487805 | |
| I0409 10:52:23.408752 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 10:52:23.408764 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.731707 | |
| I0409 10:52:23.408779 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.53801 (* 0.3 = 0.461404 loss) | |
| I0409 10:52:23.408793 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.483932 (* 0.3 = 0.14518 loss) | |
| I0409 10:52:23.408807 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.731707 | |
| I0409 10:52:23.408818 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 10:52:23.408830 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.804878 | |
| I0409 10:52:23.408844 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.851379 (* 1 = 0.851379 loss) | |
| I0409 10:52:23.408859 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.257501 (* 1 = 0.257501 loss) | |
| I0409 10:52:23.408871 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 10:52:23.408884 12249 solver.cpp:245] Train net output #16: total_confidence = 0.203123 | |
| I0409 10:52:23.408897 12249 sgd_solver.cpp:106] Iteration 143000, lr = 0.00795714 | |
| I0409 10:57:56.768846 12249 solver.cpp:229] Iteration 143500, loss = 2.4093 | |
| I0409 10:57:56.768961 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 10:57:56.768981 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 10:57:56.768995 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6 | |
| I0409 10:57:56.769011 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.21729 (* 0.3 = 0.665187 loss) | |
| I0409 10:57:56.769026 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.67497 (* 0.3 = 0.202491 loss) | |
| I0409 10:57:56.769038 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 10:57:56.769052 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 10:57:56.769063 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72 | |
| I0409 10:57:56.769078 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.5252 (* 0.3 = 0.45756 loss) | |
| I0409 10:57:56.769093 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.458934 (* 0.3 = 0.13768 loss) | |
| I0409 10:57:56.769104 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.78 | |
| I0409 10:57:56.769116 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 10:57:56.769129 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9 | |
| I0409 10:57:56.769142 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.723068 (* 1 = 0.723068 loss) | |
| I0409 10:57:56.769156 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.272653 (* 1 = 0.272653 loss) | |
| I0409 10:57:56.769168 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 10:57:56.769181 12249 solver.cpp:245] Train net output #16: total_confidence = 0.381325 | |
| I0409 10:57:56.769196 12249 sgd_solver.cpp:106] Iteration 143500, lr = 0.00795 | |
| I0409 11:03:30.174057 12249 solver.cpp:229] Iteration 144000, loss = 2.41996 | |
| I0409 11:03:30.174378 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.35 | |
| I0409 11:03:30.174408 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 11:03:30.174429 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.575 | |
| I0409 11:03:30.174458 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.12248 (* 0.3 = 0.636745 loss) | |
| I0409 11:03:30.174484 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.55918 (* 0.3 = 0.167754 loss) | |
| I0409 11:03:30.174507 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55 | |
| I0409 11:03:30.174530 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 11:03:30.174551 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.85 | |
| I0409 11:03:30.174578 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.43081 (* 0.3 = 0.429243 loss) | |
| I0409 11:03:30.174602 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.449777 (* 0.3 = 0.134933 loss) | |
| I0409 11:03:30.174628 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.825 | |
| I0409 11:03:30.174652 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 11:03:30.174675 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.925 | |
| I0409 11:03:30.174702 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.66846 (* 1 = 0.66846 loss) | |
| I0409 11:03:30.174728 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.221335 (* 1 = 0.221335 loss) | |
| I0409 11:03:30.174756 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 11:03:30.174796 12249 solver.cpp:245] Train net output #16: total_confidence = 0.439925 | |
| I0409 11:03:30.174823 12249 sgd_solver.cpp:106] Iteration 144000, lr = 0.00794286 | |
| I0409 11:09:03.517451 12249 solver.cpp:229] Iteration 144500, loss = 2.4556 | |
| I0409 11:09:03.517566 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.269231 | |
| I0409 11:09:03.517586 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 11:09:03.517599 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.576923 | |
| I0409 11:09:03.517616 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.28323 (* 0.3 = 0.684969 loss) | |
| I0409 11:09:03.517630 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.725318 (* 0.3 = 0.217595 loss) | |
| I0409 11:09:03.517643 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.461538 | |
| I0409 11:09:03.517655 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 11:09:03.517668 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.673077 | |
| I0409 11:09:03.517681 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.87124 (* 0.3 = 0.561373 loss) | |
| I0409 11:09:03.517695 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.57425 (* 0.3 = 0.172275 loss) | |
| I0409 11:09:03.517707 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.692308 | |
| I0409 11:09:03.517719 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 11:09:03.517731 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.923077 | |
| I0409 11:09:03.517748 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.06009 (* 1 = 1.06009 loss) | |
| I0409 11:09:03.517763 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.335362 (* 1 = 0.335362 loss) | |
| I0409 11:09:03.517776 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 11:09:03.517787 12249 solver.cpp:245] Train net output #16: total_confidence = 0.29044 | |
| I0409 11:09:03.517802 12249 sgd_solver.cpp:106] Iteration 144500, lr = 0.00793571 | |
| I0409 11:14:36.492895 12249 solver.cpp:338] Iteration 145000, Testing net (#0) | |
| I0409 11:15:17.931279 12249 solver.cpp:393] Test loss: 2.09945 | |
| I0409 11:15:17.931393 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.48634 | |
| I0409 11:15:17.931411 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.858048 | |
| I0409 11:15:17.931426 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.780706 | |
| I0409 11:15:17.931442 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.74367 (* 0.3 = 0.523102 loss) | |
| I0409 11:15:17.931457 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.486253 (* 0.3 = 0.145876 loss) | |
| I0409 11:15:17.931468 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.68091 | |
| I0409 11:15:17.931480 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.906548 | |
| I0409 11:15:17.931493 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.88319 | |
| I0409 11:15:17.931505 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.15371 (* 0.3 = 0.346113 loss) | |
| I0409 11:15:17.931519 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.329381 (* 0.3 = 0.0988144 loss) | |
| I0409 11:15:17.931531 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.815855 | |
| I0409 11:15:17.931543 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.951501 | |
| I0409 11:15:17.931555 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.913829 | |
| I0409 11:15:17.931568 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.777927 (* 1 = 0.777927 loss) | |
| I0409 11:15:17.931581 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.207618 (* 1 = 0.207618 loss) | |
| I0409 11:15:17.931593 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.519 | |
| I0409 11:15:17.931605 12249 solver.cpp:406] Test net output #16: total_confidence = 0.465326 | |
| I0409 11:15:18.309847 12249 solver.cpp:229] Iteration 145000, loss = 2.37244 | |
| I0409 11:15:18.309909 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.431818 | |
| I0409 11:15:18.309926 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 11:15:18.309940 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.704545 | |
| I0409 11:15:18.309957 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.02669 (* 0.3 = 0.608006 loss) | |
| I0409 11:15:18.309972 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.600878 (* 0.3 = 0.180263 loss) | |
| I0409 11:15:18.309984 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.613636 | |
| I0409 11:15:18.309998 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 11:15:18.310009 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.818182 | |
| I0409 11:15:18.310024 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.17844 (* 0.3 = 0.353533 loss) | |
| I0409 11:15:18.310039 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.376978 (* 0.3 = 0.113094 loss) | |
| I0409 11:15:18.310051 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.909091 | |
| I0409 11:15:18.310065 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 11:15:18.310076 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 11:15:18.310091 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.273503 (* 1 = 0.273503 loss) | |
| I0409 11:15:18.310106 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0910311 (* 1 = 0.0910311 loss) | |
| I0409 11:15:18.310119 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 11:15:18.310132 12249 solver.cpp:245] Train net output #16: total_confidence = 0.463848 | |
| I0409 11:15:18.310147 12249 sgd_solver.cpp:106] Iteration 145000, lr = 0.00792857 | |
| I0409 11:20:51.589089 12249 solver.cpp:229] Iteration 145500, loss = 2.39817 | |
| I0409 11:20:51.589277 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.319149 | |
| I0409 11:20:51.589298 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 11:20:51.589313 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.574468 | |
| I0409 11:20:51.589329 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.21953 (* 0.3 = 0.665858 loss) | |
| I0409 11:20:51.589344 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.644334 (* 0.3 = 0.1933 loss) | |
| I0409 11:20:51.589357 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.553191 | |
| I0409 11:20:51.589370 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 11:20:51.589382 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.702128 | |
| I0409 11:20:51.589396 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.75585 (* 0.3 = 0.526755 loss) | |
| I0409 11:20:51.589411 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.508473 (* 0.3 = 0.152542 loss) | |
| I0409 11:20:51.589422 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.595745 | |
| I0409 11:20:51.589435 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0409 11:20:51.589447 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.744681 | |
| I0409 11:20:51.589462 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.38665 (* 1 = 1.38665 loss) | |
| I0409 11:20:51.589476 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.385939 (* 1 = 0.385939 loss) | |
| I0409 11:20:51.589489 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 11:20:51.589501 12249 solver.cpp:245] Train net output #16: total_confidence = 0.295622 | |
| I0409 11:20:51.589516 12249 sgd_solver.cpp:106] Iteration 145500, lr = 0.00792143 | |
| I0409 11:26:24.960623 12249 solver.cpp:229] Iteration 146000, loss = 2.41995 | |
| I0409 11:26:24.960870 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.428571 | |
| I0409 11:26:24.960888 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 11:26:24.960901 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.642857 | |
| I0409 11:26:24.960918 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.88297 (* 0.3 = 0.56489 loss) | |
| I0409 11:26:24.960933 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.540033 (* 0.3 = 0.16201 loss) | |
| I0409 11:26:24.960945 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.595238 | |
| I0409 11:26:24.960959 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 11:26:24.960971 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.880952 | |
| I0409 11:26:24.960984 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.24416 (* 0.3 = 0.37325 loss) | |
| I0409 11:26:24.960999 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.347542 (* 0.3 = 0.104263 loss) | |
| I0409 11:26:24.961011 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.880952 | |
| I0409 11:26:24.961024 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 11:26:24.961035 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.928571 | |
| I0409 11:26:24.961050 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.463512 (* 1 = 0.463512 loss) | |
| I0409 11:26:24.961063 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.191244 (* 1 = 0.191244 loss) | |
| I0409 11:26:24.961074 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 11:26:24.961086 12249 solver.cpp:245] Train net output #16: total_confidence = 0.46006 | |
| I0409 11:26:24.961100 12249 sgd_solver.cpp:106] Iteration 146000, lr = 0.00791429 | |
| I0409 11:31:58.327209 12249 solver.cpp:229] Iteration 146500, loss = 2.36285 | |
| I0409 11:31:58.327514 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.413043 | |
| I0409 11:31:58.327535 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 11:31:58.327548 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.673913 | |
| I0409 11:31:58.327564 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.79778 (* 0.3 = 0.539335 loss) | |
| I0409 11:31:58.327580 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.546225 (* 0.3 = 0.163868 loss) | |
| I0409 11:31:58.327594 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.695652 | |
| I0409 11:31:58.327605 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 11:31:58.327617 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.891304 | |
| I0409 11:31:58.327630 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.18127 (* 0.3 = 0.354382 loss) | |
| I0409 11:31:58.327646 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.343654 (* 0.3 = 0.103096 loss) | |
| I0409 11:31:58.327657 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826087 | |
| I0409 11:31:58.327669 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 11:31:58.327682 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 11:31:58.327695 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.356999 (* 1 = 0.356999 loss) | |
| I0409 11:31:58.327709 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.124468 (* 1 = 0.124468 loss) | |
| I0409 11:31:58.327723 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 11:31:58.327733 12249 solver.cpp:245] Train net output #16: total_confidence = 0.451373 | |
| I0409 11:31:58.327750 12249 sgd_solver.cpp:106] Iteration 146500, lr = 0.00790714 | |
| I0409 11:37:31.703477 12249 solver.cpp:229] Iteration 147000, loss = 2.41608 | |
| I0409 11:37:31.703585 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.463415 | |
| I0409 11:37:31.703604 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.869318 | |
| I0409 11:37:31.703618 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.756098 | |
| I0409 11:37:31.703634 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.86077 (* 0.3 = 0.55823 loss) | |
| I0409 11:37:31.703649 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.50595 (* 0.3 = 0.151785 loss) | |
| I0409 11:37:31.703662 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.609756 | |
| I0409 11:37:31.703675 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 11:37:31.703686 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.853659 | |
| I0409 11:37:31.703701 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.2391 (* 0.3 = 0.371729 loss) | |
| I0409 11:37:31.703716 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.345398 (* 0.3 = 0.10362 loss) | |
| I0409 11:37:31.703727 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.756098 | |
| I0409 11:37:31.703739 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 11:37:31.703754 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.926829 | |
| I0409 11:37:31.703769 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.728068 (* 1 = 0.728068 loss) | |
| I0409 11:37:31.703783 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.199975 (* 1 = 0.199975 loss) | |
| I0409 11:37:31.703796 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 11:37:31.703809 12249 solver.cpp:245] Train net output #16: total_confidence = 0.374089 | |
| I0409 11:37:31.703824 12249 sgd_solver.cpp:106] Iteration 147000, lr = 0.0079 | |
| I0409 11:43:05.064947 12249 solver.cpp:229] Iteration 147500, loss = 2.39906 | |
| I0409 11:43:05.065258 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.232558 | |
| I0409 11:43:05.065279 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 11:43:05.065292 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.372093 | |
| I0409 11:43:05.065310 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.71933 (* 0.3 = 0.8158 loss) | |
| I0409 11:43:05.065325 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.814624 (* 0.3 = 0.244387 loss) | |
| I0409 11:43:05.065337 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.465116 | |
| I0409 11:43:05.065349 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 11:43:05.065361 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72093 | |
| I0409 11:43:05.065376 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.84638 (* 0.3 = 0.553914 loss) | |
| I0409 11:43:05.065389 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.650321 (* 0.3 = 0.195096 loss) | |
| I0409 11:43:05.065402 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.790698 | |
| I0409 11:43:05.065415 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 11:43:05.065428 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.883721 | |
| I0409 11:43:05.065441 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.784113 (* 1 = 0.784113 loss) | |
| I0409 11:43:05.065455 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.203787 (* 1 = 0.203787 loss) | |
| I0409 11:43:05.065469 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 11:43:05.065480 12249 solver.cpp:245] Train net output #16: total_confidence = 0.409508 | |
| I0409 11:43:05.065495 12249 sgd_solver.cpp:106] Iteration 147500, lr = 0.00789286 | |
| I0409 11:48:38.435695 12249 solver.cpp:229] Iteration 148000, loss = 2.29894 | |
| I0409 11:48:38.435817 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.326923 | |
| I0409 11:48:38.435837 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 11:48:38.435850 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.576923 | |
| I0409 11:48:38.435868 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.2238 (* 0.3 = 0.667139 loss) | |
| I0409 11:48:38.435883 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.728121 (* 0.3 = 0.218436 loss) | |
| I0409 11:48:38.435895 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.596154 | |
| I0409 11:48:38.435909 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 11:48:38.435920 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.826923 | |
| I0409 11:48:38.435935 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.63736 (* 0.3 = 0.491207 loss) | |
| I0409 11:48:38.435950 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.504454 (* 0.3 = 0.151336 loss) | |
| I0409 11:48:38.435961 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.807692 | |
| I0409 11:48:38.435974 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 11:48:38.435986 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.923077 | |
| I0409 11:48:38.436002 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.786096 (* 1 = 0.786096 loss) | |
| I0409 11:48:38.436015 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.236743 (* 1 = 0.236743 loss) | |
| I0409 11:48:38.436028 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 11:48:38.436040 12249 solver.cpp:245] Train net output #16: total_confidence = 0.403394 | |
| I0409 11:48:38.436054 12249 sgd_solver.cpp:106] Iteration 148000, lr = 0.00788571 | |
| I0409 11:54:12.145874 12249 solver.cpp:229] Iteration 148500, loss = 2.32786 | |
| I0409 11:54:12.146205 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.439024 | |
| I0409 11:54:12.146226 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 11:54:12.146239 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.707317 | |
| I0409 11:54:12.146256 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.81257 (* 0.3 = 0.54377 loss) | |
| I0409 11:54:12.146271 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.496792 (* 0.3 = 0.149038 loss) | |
| I0409 11:54:12.146284 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.707317 | |
| I0409 11:54:12.146296 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 11:54:12.146308 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.902439 | |
| I0409 11:54:12.146322 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.11566 (* 0.3 = 0.334697 loss) | |
| I0409 11:54:12.146337 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.340156 (* 0.3 = 0.102047 loss) | |
| I0409 11:54:12.146348 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.878049 | |
| I0409 11:54:12.146360 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 11:54:12.146373 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.97561 | |
| I0409 11:54:12.146387 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.421557 (* 1 = 0.421557 loss) | |
| I0409 11:54:12.146401 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.113659 (* 1 = 0.113659 loss) | |
| I0409 11:54:12.146414 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 11:54:12.146426 12249 solver.cpp:245] Train net output #16: total_confidence = 0.509454 | |
| I0409 11:54:12.146441 12249 sgd_solver.cpp:106] Iteration 148500, lr = 0.00787857 | |
| I0409 11:59:45.513896 12249 solver.cpp:229] Iteration 149000, loss = 2.36902 | |
| I0409 11:59:45.514014 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 11:59:45.514034 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 11:59:45.514046 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.725 | |
| I0409 11:59:45.514062 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.94374 (* 0.3 = 0.583122 loss) | |
| I0409 11:59:45.514077 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.499473 (* 0.3 = 0.149842 loss) | |
| I0409 11:59:45.514091 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.575 | |
| I0409 11:59:45.514103 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 11:59:45.514116 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.85 | |
| I0409 11:59:45.514128 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.36714 (* 0.3 = 0.410141 loss) | |
| I0409 11:59:45.514143 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.381281 (* 0.3 = 0.114384 loss) | |
| I0409 11:59:45.514155 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.9 | |
| I0409 11:59:45.514168 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 11:59:45.514179 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.95 | |
| I0409 11:59:45.514194 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.695352 (* 1 = 0.695352 loss) | |
| I0409 11:59:45.514207 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.160875 (* 1 = 0.160875 loss) | |
| I0409 11:59:45.514219 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 11:59:45.514232 12249 solver.cpp:245] Train net output #16: total_confidence = 0.544589 | |
| I0409 11:59:45.514246 12249 sgd_solver.cpp:106] Iteration 149000, lr = 0.00787143 | |
| I0409 12:05:18.878000 12249 solver.cpp:229] Iteration 149500, loss = 2.34995 | |
| I0409 12:05:18.878314 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 12:05:18.878334 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 12:05:18.878347 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.64 | |
| I0409 12:05:18.878365 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.90832 (* 0.3 = 0.572495 loss) | |
| I0409 12:05:18.878379 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.612451 (* 0.3 = 0.183735 loss) | |
| I0409 12:05:18.878392 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.52 | |
| I0409 12:05:18.878404 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 12:05:18.878417 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.74 | |
| I0409 12:05:18.878432 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52321 (* 0.3 = 0.456963 loss) | |
| I0409 12:05:18.878445 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.469903 (* 0.3 = 0.140971 loss) | |
| I0409 12:05:18.878458 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.78 | |
| I0409 12:05:18.878470 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 12:05:18.878482 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.9 | |
| I0409 12:05:18.878497 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.608615 (* 1 = 0.608615 loss) | |
| I0409 12:05:18.878511 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.182256 (* 1 = 0.182256 loss) | |
| I0409 12:05:18.878525 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 12:05:18.878536 12249 solver.cpp:245] Train net output #16: total_confidence = 0.404554 | |
| I0409 12:05:18.878551 12249 sgd_solver.cpp:106] Iteration 149500, lr = 0.00786429 | |
| I0409 12:10:51.857184 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_150000.caffemodel | |
| I0409 12:10:52.340406 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_150000.solverstate | |
| I0409 12:10:52.584153 12249 solver.cpp:338] Iteration 150000, Testing net (#0) | |
| I0409 12:11:33.607570 12249 solver.cpp:393] Test loss: 2.13854 | |
| I0409 12:11:33.607830 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.438572 | |
| I0409 12:11:33.607848 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.856457 | |
| I0409 12:11:33.607861 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.749453 | |
| I0409 12:11:33.607877 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.83468 (* 0.3 = 0.550405 loss) | |
| I0409 12:11:33.607892 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.477656 (* 0.3 = 0.143297 loss) | |
| I0409 12:11:33.607903 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.674428 | |
| I0409 12:11:33.607915 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.91582 | |
| I0409 12:11:33.607928 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.877198 | |
| I0409 12:11:33.607940 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.18024 (* 0.3 = 0.354073 loss) | |
| I0409 12:11:33.607954 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.309566 (* 0.3 = 0.0928698 loss) | |
| I0409 12:11:33.607966 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.813568 | |
| I0409 12:11:33.607978 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.954318 | |
| I0409 12:11:33.607990 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.90797 | |
| I0409 12:11:33.608003 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.795938 (* 1 = 0.795938 loss) | |
| I0409 12:11:33.608016 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.201956 (* 1 = 0.201956 loss) | |
| I0409 12:11:33.608028 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.555 | |
| I0409 12:11:33.608039 12249 solver.cpp:406] Test net output #16: total_confidence = 0.46665 | |
| I0409 12:11:33.980741 12249 solver.cpp:229] Iteration 150000, loss = 2.38211 | |
| I0409 12:11:33.980800 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.568182 | |
| I0409 12:11:33.980818 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.869318 | |
| I0409 12:11:33.980831 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.840909 | |
| I0409 12:11:33.980849 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.48624 (* 0.3 = 0.445873 loss) | |
| I0409 12:11:33.980864 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.452518 (* 0.3 = 0.135755 loss) | |
| I0409 12:11:33.980875 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.704545 | |
| I0409 12:11:33.980888 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 12:11:33.980901 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.909091 | |
| I0409 12:11:33.980914 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.00486 (* 0.3 = 0.301457 loss) | |
| I0409 12:11:33.980929 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.329121 (* 0.3 = 0.0987364 loss) | |
| I0409 12:11:33.980942 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.931818 | |
| I0409 12:11:33.980954 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 12:11:33.980967 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 12:11:33.980981 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.271672 (* 1 = 0.271672 loss) | |
| I0409 12:11:33.980995 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.116727 (* 1 = 0.116727 loss) | |
| I0409 12:11:33.981009 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 12:11:33.981020 12249 solver.cpp:245] Train net output #16: total_confidence = 0.540281 | |
| I0409 12:11:33.981035 12249 sgd_solver.cpp:106] Iteration 150000, lr = 0.00785714 | |
| I0409 12:17:07.293083 12249 solver.cpp:229] Iteration 150500, loss = 2.37127 | |
| I0409 12:17:07.293243 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.372549 | |
| I0409 12:17:07.293272 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 12:17:07.293298 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.666667 | |
| I0409 12:17:07.293330 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.99868 (* 0.3 = 0.599603 loss) | |
| I0409 12:17:07.293364 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.621207 (* 0.3 = 0.186362 loss) | |
| I0409 12:17:07.293395 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.529412 | |
| I0409 12:17:07.293422 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 12:17:07.293439 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.843137 | |
| I0409 12:17:07.293453 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.40565 (* 0.3 = 0.421694 loss) | |
| I0409 12:17:07.293468 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.471031 (* 0.3 = 0.141309 loss) | |
| I0409 12:17:07.293480 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.784314 | |
| I0409 12:17:07.293493 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 12:17:07.293505 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.882353 | |
| I0409 12:17:07.293519 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.732834 (* 1 = 0.732834 loss) | |
| I0409 12:17:07.293534 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.251158 (* 1 = 0.251158 loss) | |
| I0409 12:17:07.293545 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 12:17:07.293558 12249 solver.cpp:245] Train net output #16: total_confidence = 0.397505 | |
| I0409 12:17:07.293572 12249 sgd_solver.cpp:106] Iteration 150500, lr = 0.00785 | |
| I0409 12:22:40.664530 12249 solver.cpp:229] Iteration 151000, loss = 2.41054 | |
| I0409 12:22:40.664829 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.238095 | |
| I0409 12:22:40.664850 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 12:22:40.664865 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.547619 | |
| I0409 12:22:40.664880 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.48105 (* 0.3 = 0.744316 loss) | |
| I0409 12:22:40.664896 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.685568 (* 0.3 = 0.20567 loss) | |
| I0409 12:22:40.664908 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 12:22:40.664921 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 12:22:40.664932 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.738095 | |
| I0409 12:22:40.664947 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.46689 (* 0.3 = 0.440067 loss) | |
| I0409 12:22:40.664961 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.50973 (* 0.3 = 0.152919 loss) | |
| I0409 12:22:40.664974 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.833333 | |
| I0409 12:22:40.664986 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 12:22:40.664997 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.928571 | |
| I0409 12:22:40.665012 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.799514 (* 1 = 0.799514 loss) | |
| I0409 12:22:40.665026 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.210474 (* 1 = 0.210474 loss) | |
| I0409 12:22:40.665038 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 12:22:40.665051 12249 solver.cpp:245] Train net output #16: total_confidence = 0.256115 | |
| I0409 12:22:40.665066 12249 sgd_solver.cpp:106] Iteration 151000, lr = 0.00784286 | |
| I0409 12:28:14.031847 12249 solver.cpp:229] Iteration 151500, loss = 2.36512 | |
| I0409 12:28:14.031980 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 12:28:14.032002 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 12:28:14.032016 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.644444 | |
| I0409 12:28:14.032032 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.26751 (* 0.3 = 0.680254 loss) | |
| I0409 12:28:14.032048 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.646487 (* 0.3 = 0.193946 loss) | |
| I0409 12:28:14.032061 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.6 | |
| I0409 12:28:14.032073 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 12:28:14.032085 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.822222 | |
| I0409 12:28:14.032099 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.31746 (* 0.3 = 0.395237 loss) | |
| I0409 12:28:14.032114 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.450885 (* 0.3 = 0.135265 loss) | |
| I0409 12:28:14.032127 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.822222 | |
| I0409 12:28:14.032140 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 12:28:14.032151 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.933333 | |
| I0409 12:28:14.032166 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.596286 (* 1 = 0.596286 loss) | |
| I0409 12:28:14.032181 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.212457 (* 1 = 0.212457 loss) | |
| I0409 12:28:14.032192 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 12:28:14.032205 12249 solver.cpp:245] Train net output #16: total_confidence = 0.352042 | |
| I0409 12:28:14.032220 12249 sgd_solver.cpp:106] Iteration 151500, lr = 0.00783571 | |
| I0409 12:33:47.406499 12249 solver.cpp:229] Iteration 152000, loss = 2.3002 | |
| I0409 12:33:47.406831 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.5 | |
| I0409 12:33:47.406852 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 12:33:47.406867 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.785714 | |
| I0409 12:33:47.406883 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.62707 (* 0.3 = 0.488122 loss) | |
| I0409 12:33:47.406898 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.463339 (* 0.3 = 0.139002 loss) | |
| I0409 12:33:47.406911 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.809524 | |
| I0409 12:33:47.406924 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.926136 | |
| I0409 12:33:47.406935 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.904762 | |
| I0409 12:33:47.406949 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.09297 (* 0.3 = 0.32789 loss) | |
| I0409 12:33:47.406965 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.328621 (* 0.3 = 0.0985862 loss) | |
| I0409 12:33:47.406977 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.857143 | |
| I0409 12:33:47.406990 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 12:33:47.407001 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.928571 | |
| I0409 12:33:47.407016 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.559465 (* 1 = 0.559465 loss) | |
| I0409 12:33:47.407029 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.160635 (* 1 = 0.160635 loss) | |
| I0409 12:33:47.407042 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 12:33:47.407053 12249 solver.cpp:245] Train net output #16: total_confidence = 0.631278 | |
| I0409 12:33:47.407068 12249 sgd_solver.cpp:106] Iteration 152000, lr = 0.00782857 | |
| I0409 12:39:20.776233 12249 solver.cpp:229] Iteration 152500, loss = 2.34007 | |
| I0409 12:39:20.776345 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.488889 | |
| I0409 12:39:20.776365 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 12:39:20.776379 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.777778 | |
| I0409 12:39:20.776396 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.80512 (* 0.3 = 0.541537 loss) | |
| I0409 12:39:20.776410 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.495624 (* 0.3 = 0.148687 loss) | |
| I0409 12:39:20.776423 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.577778 | |
| I0409 12:39:20.776437 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 12:39:20.776448 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.911111 | |
| I0409 12:39:20.776463 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.25169 (* 0.3 = 0.375508 loss) | |
| I0409 12:39:20.776476 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.386146 (* 0.3 = 0.115844 loss) | |
| I0409 12:39:20.776504 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.866667 | |
| I0409 12:39:20.776517 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 12:39:20.776530 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 12:39:20.776545 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.324612 (* 1 = 0.324612 loss) | |
| I0409 12:39:20.776569 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0883524 (* 1 = 0.0883524 loss) | |
| I0409 12:39:20.776588 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 12:39:20.776607 12249 solver.cpp:245] Train net output #16: total_confidence = 0.439943 | |
| I0409 12:39:20.776621 12249 sgd_solver.cpp:106] Iteration 152500, lr = 0.00782143 | |
| I0409 12:44:54.803848 12249 solver.cpp:229] Iteration 153000, loss = 2.29679 | |
| I0409 12:44:54.804143 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.416667 | |
| I0409 12:44:54.804164 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 12:44:54.804177 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.708333 | |
| I0409 12:44:54.804193 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.94989 (* 0.3 = 0.584968 loss) | |
| I0409 12:44:54.804208 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.660749 (* 0.3 = 0.198225 loss) | |
| I0409 12:44:54.804221 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.583333 | |
| I0409 12:44:54.804234 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 12:44:54.804245 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.8125 | |
| I0409 12:44:54.804260 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.45164 (* 0.3 = 0.435493 loss) | |
| I0409 12:44:54.804273 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.513607 (* 0.3 = 0.154082 loss) | |
| I0409 12:44:54.804286 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.770833 | |
| I0409 12:44:54.804297 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 12:44:54.804309 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.916667 | |
| I0409 12:44:54.804323 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.691592 (* 1 = 0.691592 loss) | |
| I0409 12:44:54.804337 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.243755 (* 1 = 0.243755 loss) | |
| I0409 12:44:54.804349 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 12:44:54.804361 12249 solver.cpp:245] Train net output #16: total_confidence = 0.368882 | |
| I0409 12:44:54.804376 12249 sgd_solver.cpp:106] Iteration 153000, lr = 0.00781429 | |
| I0409 12:50:28.187567 12249 solver.cpp:229] Iteration 153500, loss = 2.32309 | |
| I0409 12:50:28.187695 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.352941 | |
| I0409 12:50:28.187714 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 12:50:28.187727 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.588235 | |
| I0409 12:50:28.187747 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.30456 (* 0.3 = 0.691368 loss) | |
| I0409 12:50:28.187762 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.713087 (* 0.3 = 0.213926 loss) | |
| I0409 12:50:28.187775 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.411765 | |
| I0409 12:50:28.187788 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 12:50:28.187799 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.862745 | |
| I0409 12:50:28.187814 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.55174 (* 0.3 = 0.465523 loss) | |
| I0409 12:50:28.187829 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.468545 (* 0.3 = 0.140564 loss) | |
| I0409 12:50:28.187840 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.784314 | |
| I0409 12:50:28.187852 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 12:50:28.187865 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.980392 | |
| I0409 12:50:28.187880 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.614926 (* 1 = 0.614926 loss) | |
| I0409 12:50:28.187893 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.18066 (* 1 = 0.18066 loss) | |
| I0409 12:50:28.187906 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 12:50:28.187918 12249 solver.cpp:245] Train net output #16: total_confidence = 0.45317 | |
| I0409 12:50:28.187933 12249 sgd_solver.cpp:106] Iteration 153500, lr = 0.00780714 | |
| I0409 12:56:01.546021 12249 solver.cpp:229] Iteration 154000, loss = 2.33782 | |
| I0409 12:56:01.546385 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.2 | |
| I0409 12:56:01.546406 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 12:56:01.546421 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.577778 | |
| I0409 12:56:01.546437 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.73017 (* 0.3 = 0.819051 loss) | |
| I0409 12:56:01.546452 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.757373 (* 0.3 = 0.227212 loss) | |
| I0409 12:56:01.546464 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4 | |
| I0409 12:56:01.546478 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 12:56:01.546489 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.666667 | |
| I0409 12:56:01.546505 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.13765 (* 0.3 = 0.641295 loss) | |
| I0409 12:56:01.546519 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.669335 (* 0.3 = 0.2008 loss) | |
| I0409 12:56:01.546531 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.711111 | |
| I0409 12:56:01.546545 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 12:56:01.546556 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.844444 | |
| I0409 12:56:01.546571 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.07537 (* 1 = 1.07537 loss) | |
| I0409 12:56:01.546584 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.326845 (* 1 = 0.326845 loss) | |
| I0409 12:56:01.546597 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 12:56:01.546609 12249 solver.cpp:245] Train net output #16: total_confidence = 0.241872 | |
| I0409 12:56:01.546624 12249 sgd_solver.cpp:106] Iteration 154000, lr = 0.0078 | |
| I0409 13:01:34.922852 12249 solver.cpp:229] Iteration 154500, loss = 2.33109 | |
| I0409 13:01:34.923153 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.44 | |
| I0409 13:01:34.923176 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 13:01:34.923189 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.58 | |
| I0409 13:01:34.923205 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.51419 (* 0.3 = 0.754257 loss) | |
| I0409 13:01:34.923220 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.758725 (* 0.3 = 0.227618 loss) | |
| I0409 13:01:34.923233 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.4 | |
| I0409 13:01:34.923245 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 13:01:34.923257 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.74 | |
| I0409 13:01:34.923271 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.86883 (* 0.3 = 0.56065 loss) | |
| I0409 13:01:34.923285 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.5635 (* 0.3 = 0.16905 loss) | |
| I0409 13:01:34.923298 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.74 | |
| I0409 13:01:34.923310 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 13:01:34.923322 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.88 | |
| I0409 13:01:34.923337 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.933676 (* 1 = 0.933676 loss) | |
| I0409 13:01:34.923352 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.300974 (* 1 = 0.300974 loss) | |
| I0409 13:01:34.923365 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 13:01:34.923378 12249 solver.cpp:245] Train net output #16: total_confidence = 0.307397 | |
| I0409 13:01:34.923393 12249 sgd_solver.cpp:106] Iteration 154500, lr = 0.00779286 | |
| I0409 13:07:08.557943 12249 solver.cpp:338] Iteration 155000, Testing net (#0) | |
| I0409 13:07:49.512954 12249 solver.cpp:393] Test loss: 2.14572 | |
| I0409 13:07:49.513067 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.446503 | |
| I0409 13:07:49.513087 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.859503 | |
| I0409 13:07:49.513099 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.752926 | |
| I0409 13:07:49.513115 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.90033 (* 0.3 = 0.570098 loss) | |
| I0409 13:07:49.513130 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.488796 (* 0.3 = 0.146639 loss) | |
| I0409 13:07:49.513142 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.683821 | |
| I0409 13:07:49.513154 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.915182 | |
| I0409 13:07:49.513165 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.892571 | |
| I0409 13:07:49.513180 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.13327 (* 0.3 = 0.339982 loss) | |
| I0409 13:07:49.513193 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.303137 (* 0.3 = 0.090941 loss) | |
| I0409 13:07:49.513206 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.817207 | |
| I0409 13:07:49.513217 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.955319 | |
| I0409 13:07:49.513229 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.913199 | |
| I0409 13:07:49.513243 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.7953 (* 1 = 0.7953 loss) | |
| I0409 13:07:49.513257 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.202761 (* 1 = 0.202761 loss) | |
| I0409 13:07:49.513268 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.572 | |
| I0409 13:07:49.513280 12249 solver.cpp:406] Test net output #16: total_confidence = 0.520902 | |
| I0409 13:07:49.885273 12249 solver.cpp:229] Iteration 155000, loss = 2.29698 | |
| I0409 13:07:49.885324 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.28 | |
| I0409 13:07:49.885341 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 13:07:49.885354 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.66 | |
| I0409 13:07:49.885371 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.15119 (* 0.3 = 0.645357 loss) | |
| I0409 13:07:49.885386 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.655449 (* 0.3 = 0.196635 loss) | |
| I0409 13:07:49.885399 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.54 | |
| I0409 13:07:49.885412 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 13:07:49.885424 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.82 | |
| I0409 13:07:49.885438 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.41739 (* 0.3 = 0.425218 loss) | |
| I0409 13:07:49.885452 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.434802 (* 0.3 = 0.130441 loss) | |
| I0409 13:07:49.885465 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.82 | |
| I0409 13:07:49.885478 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 13:07:49.885489 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.96 | |
| I0409 13:07:49.885504 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.550752 (* 1 = 0.550752 loss) | |
| I0409 13:07:49.885517 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.162076 (* 1 = 0.162076 loss) | |
| I0409 13:07:49.885530 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 13:07:49.885542 12249 solver.cpp:245] Train net output #16: total_confidence = 0.411567 | |
| I0409 13:07:49.885556 12249 sgd_solver.cpp:106] Iteration 155000, lr = 0.00778571 | |
| I0409 13:13:23.334980 12249 solver.cpp:229] Iteration 155500, loss = 2.33078 | |
| I0409 13:13:23.335392 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.358491 | |
| I0409 13:13:23.335414 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 13:13:23.335428 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.603774 | |
| I0409 13:13:23.335445 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.26151 (* 0.3 = 0.678452 loss) | |
| I0409 13:13:23.335460 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.699804 (* 0.3 = 0.209941 loss) | |
| I0409 13:13:23.335474 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.45283 | |
| I0409 13:13:23.335486 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 13:13:23.335499 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.811321 | |
| I0409 13:13:23.335512 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.4644 (* 0.3 = 0.439319 loss) | |
| I0409 13:13:23.335527 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.458808 (* 0.3 = 0.137643 loss) | |
| I0409 13:13:23.335541 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.830189 | |
| I0409 13:13:23.335552 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 13:13:23.335564 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.943396 | |
| I0409 13:13:23.335579 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.661218 (* 1 = 0.661218 loss) | |
| I0409 13:13:23.335593 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.220017 (* 1 = 0.220017 loss) | |
| I0409 13:13:23.335607 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 13:13:23.335618 12249 solver.cpp:245] Train net output #16: total_confidence = 0.401831 | |
| I0409 13:13:23.335633 12249 sgd_solver.cpp:106] Iteration 155500, lr = 0.00777857 | |
| I0409 13:18:56.694002 12249 solver.cpp:229] Iteration 156000, loss = 2.33146 | |
| I0409 13:18:56.694124 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.511628 | |
| I0409 13:18:56.694144 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.875 | |
| I0409 13:18:56.694156 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.837209 | |
| I0409 13:18:56.694174 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.52479 (* 0.3 = 0.457436 loss) | |
| I0409 13:18:56.694190 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.415702 (* 0.3 = 0.124711 loss) | |
| I0409 13:18:56.694202 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.604651 | |
| I0409 13:18:56.694216 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 13:18:56.694227 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.860465 | |
| I0409 13:18:56.694242 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.35108 (* 0.3 = 0.405323 loss) | |
| I0409 13:18:56.694255 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.383366 (* 0.3 = 0.11501 loss) | |
| I0409 13:18:56.694268 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.860465 | |
| I0409 13:18:56.694280 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 13:18:56.694293 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.930233 | |
| I0409 13:18:56.694306 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.655119 (* 1 = 0.655119 loss) | |
| I0409 13:18:56.694320 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.164986 (* 1 = 0.164986 loss) | |
| I0409 13:18:56.694334 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 13:18:56.694345 12249 solver.cpp:245] Train net output #16: total_confidence = 0.620643 | |
| I0409 13:18:56.694360 12249 sgd_solver.cpp:106] Iteration 156000, lr = 0.00777143 | |
| I0409 13:24:30.840188 12249 solver.cpp:229] Iteration 156500, loss = 2.28411 | |
| I0409 13:24:30.840657 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.294118 | |
| I0409 13:24:30.840679 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 13:24:30.840693 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.568627 | |
| I0409 13:24:30.840709 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.68535 (* 0.3 = 0.805605 loss) | |
| I0409 13:24:30.840725 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.850856 (* 0.3 = 0.255257 loss) | |
| I0409 13:24:30.840737 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.431373 | |
| I0409 13:24:30.840754 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.806818 | |
| I0409 13:24:30.840766 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.705882 | |
| I0409 13:24:30.840780 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.07506 (* 0.3 = 0.622517 loss) | |
| I0409 13:24:30.840795 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.674968 (* 0.3 = 0.20249 loss) | |
| I0409 13:24:30.840807 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.627451 | |
| I0409 13:24:30.840819 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.869318 | |
| I0409 13:24:30.840831 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.72549 | |
| I0409 13:24:30.840847 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.31738 (* 1 = 2.31738 loss) | |
| I0409 13:24:30.840860 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.749407 (* 1 = 0.749407 loss) | |
| I0409 13:24:30.840873 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 13:24:30.840884 12249 solver.cpp:245] Train net output #16: total_confidence = 0.311255 | |
| I0409 13:24:30.840899 12249 sgd_solver.cpp:106] Iteration 156500, lr = 0.00776429 | |
| I0409 13:30:04.113538 12249 solver.cpp:229] Iteration 157000, loss = 2.3113 | |
| I0409 13:30:04.113661 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 13:30:04.113680 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 13:30:04.113693 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.62 | |
| I0409 13:30:04.113710 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.01085 (* 0.3 = 0.603254 loss) | |
| I0409 13:30:04.113725 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.626387 (* 0.3 = 0.187916 loss) | |
| I0409 13:30:04.113739 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 13:30:04.113754 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 13:30:04.113766 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.82 | |
| I0409 13:30:04.113780 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.41666 (* 0.3 = 0.424997 loss) | |
| I0409 13:30:04.113795 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.429333 (* 0.3 = 0.1288 loss) | |
| I0409 13:30:04.113807 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.8 | |
| I0409 13:30:04.113819 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 13:30:04.113831 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.98 | |
| I0409 13:30:04.113847 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.544056 (* 1 = 0.544056 loss) | |
| I0409 13:30:04.113862 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.164404 (* 1 = 0.164404 loss) | |
| I0409 13:30:04.113873 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 13:30:04.113885 12249 solver.cpp:245] Train net output #16: total_confidence = 0.343418 | |
| I0409 13:30:04.113899 12249 sgd_solver.cpp:106] Iteration 157000, lr = 0.00775714 | |
| I0409 13:35:37.490679 12249 solver.cpp:229] Iteration 157500, loss = 2.29463 | |
| I0409 13:35:37.490984 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.404255 | |
| I0409 13:35:37.491005 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 13:35:37.491019 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.595745 | |
| I0409 13:35:37.491035 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.08247 (* 0.3 = 0.624742 loss) | |
| I0409 13:35:37.491050 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.634509 (* 0.3 = 0.190353 loss) | |
| I0409 13:35:37.491063 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.553191 | |
| I0409 13:35:37.491076 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 13:35:37.491088 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.87234 | |
| I0409 13:35:37.491101 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.26563 (* 0.3 = 0.379689 loss) | |
| I0409 13:35:37.491117 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.385916 (* 0.3 = 0.115775 loss) | |
| I0409 13:35:37.491129 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.87234 | |
| I0409 13:35:37.491142 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 13:35:37.491154 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0409 13:35:37.491168 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.512256 (* 1 = 0.512256 loss) | |
| I0409 13:35:37.491183 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.158955 (* 1 = 0.158955 loss) | |
| I0409 13:35:37.491194 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 13:35:37.491206 12249 solver.cpp:245] Train net output #16: total_confidence = 0.287908 | |
| I0409 13:35:37.491221 12249 sgd_solver.cpp:106] Iteration 157500, lr = 0.00775 | |
| I0409 13:41:10.856710 12249 solver.cpp:229] Iteration 158000, loss = 2.37219 | |
| I0409 13:41:10.856843 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.444444 | |
| I0409 13:41:10.856864 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 13:41:10.856878 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.733333 | |
| I0409 13:41:10.856894 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.70456 (* 0.3 = 0.511367 loss) | |
| I0409 13:41:10.856909 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.510376 (* 0.3 = 0.153113 loss) | |
| I0409 13:41:10.856922 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.666667 | |
| I0409 13:41:10.856935 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 13:41:10.856946 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.844444 | |
| I0409 13:41:10.856961 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.12424 (* 0.3 = 0.337272 loss) | |
| I0409 13:41:10.856974 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.328838 (* 0.3 = 0.0986513 loss) | |
| I0409 13:41:10.856987 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.888889 | |
| I0409 13:41:10.856999 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 13:41:10.857012 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.911111 | |
| I0409 13:41:10.857025 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.495001 (* 1 = 0.495001 loss) | |
| I0409 13:41:10.857039 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.152968 (* 1 = 0.152968 loss) | |
| I0409 13:41:10.857051 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 13:41:10.857064 12249 solver.cpp:245] Train net output #16: total_confidence = 0.604225 | |
| I0409 13:41:10.857079 12249 sgd_solver.cpp:106] Iteration 158000, lr = 0.00774286 | |
| I0409 13:46:44.222134 12249 solver.cpp:229] Iteration 158500, loss = 2.28311 | |
| I0409 13:46:44.222426 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.361702 | |
| I0409 13:46:44.222448 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 13:46:44.222462 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.659574 | |
| I0409 13:46:44.222478 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95487 (* 0.3 = 0.586461 loss) | |
| I0409 13:46:44.222494 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.583493 (* 0.3 = 0.175048 loss) | |
| I0409 13:46:44.222506 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.468085 | |
| I0409 13:46:44.222519 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 13:46:44.222532 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.93617 | |
| I0409 13:46:44.222544 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.28939 (* 0.3 = 0.386818 loss) | |
| I0409 13:46:44.222559 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.381496 (* 0.3 = 0.114449 loss) | |
| I0409 13:46:44.222571 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.787234 | |
| I0409 13:46:44.222584 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 13:46:44.222595 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.93617 | |
| I0409 13:46:44.222610 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.847494 (* 1 = 0.847494 loss) | |
| I0409 13:46:44.222625 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.244825 (* 1 = 0.244825 loss) | |
| I0409 13:46:44.222636 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 13:46:44.222648 12249 solver.cpp:245] Train net output #16: total_confidence = 0.409671 | |
| I0409 13:46:44.222663 12249 sgd_solver.cpp:106] Iteration 158500, lr = 0.00773571 | |
| I0409 13:52:17.608281 12249 solver.cpp:229] Iteration 159000, loss = 2.33322 | |
| I0409 13:52:17.608638 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.416667 | |
| I0409 13:52:17.608659 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 13:52:17.608672 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.645833 | |
| I0409 13:52:17.608690 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.08109 (* 0.3 = 0.624326 loss) | |
| I0409 13:52:17.608705 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.647039 (* 0.3 = 0.194112 loss) | |
| I0409 13:52:17.608717 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.604167 | |
| I0409 13:52:17.608731 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 13:52:17.608744 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.8125 | |
| I0409 13:52:17.608760 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.16751 (* 0.3 = 0.650252 loss) | |
| I0409 13:52:17.608774 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.665714 (* 0.3 = 0.199714 loss) | |
| I0409 13:52:17.608788 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.770833 | |
| I0409 13:52:17.608799 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 13:52:17.608811 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.875 | |
| I0409 13:52:17.608825 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.846451 (* 1 = 0.846451 loss) | |
| I0409 13:52:17.608839 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.242464 (* 1 = 0.242464 loss) | |
| I0409 13:52:17.608852 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 13:52:17.608865 12249 solver.cpp:245] Train net output #16: total_confidence = 0.476819 | |
| I0409 13:52:17.608880 12249 sgd_solver.cpp:106] Iteration 159000, lr = 0.00772857 | |
| I0409 13:57:50.968025 12249 solver.cpp:229] Iteration 159500, loss = 2.32734 | |
| I0409 13:57:50.968258 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.564103 | |
| I0409 13:57:50.968279 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.880682 | |
| I0409 13:57:50.968293 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.846154 | |
| I0409 13:57:50.968310 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.64598 (* 0.3 = 0.493794 loss) | |
| I0409 13:57:50.968325 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.447389 (* 0.3 = 0.134217 loss) | |
| I0409 13:57:50.968338 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.717949 | |
| I0409 13:57:50.968351 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.914773 | |
| I0409 13:57:50.968364 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.846154 | |
| I0409 13:57:50.968377 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.11638 (* 0.3 = 0.334913 loss) | |
| I0409 13:57:50.968392 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.304046 (* 0.3 = 0.0912137 loss) | |
| I0409 13:57:50.968405 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.846154 | |
| I0409 13:57:50.968417 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 13:57:50.968430 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.923077 | |
| I0409 13:57:50.968443 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.440641 (* 1 = 0.440641 loss) | |
| I0409 13:57:50.968458 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.10758 (* 1 = 0.10758 loss) | |
| I0409 13:57:50.968472 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 13:57:50.968497 12249 solver.cpp:245] Train net output #16: total_confidence = 0.508475 | |
| I0409 13:57:50.968514 12249 sgd_solver.cpp:106] Iteration 159500, lr = 0.00772143 | |
| I0409 14:03:23.956611 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_160000.caffemodel | |
| I0409 14:03:24.438662 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_160000.solverstate | |
| I0409 14:03:24.685377 12249 solver.cpp:338] Iteration 160000, Testing net (#0) | |
| I0409 14:04:05.828133 12249 solver.cpp:393] Test loss: 2.26081 | |
| I0409 14:04:05.828280 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.461289 | |
| I0409 14:04:05.828299 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.859276 | |
| I0409 14:04:05.828313 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.721346 | |
| I0409 14:04:05.828330 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.94924 (* 0.3 = 0.584771 loss) | |
| I0409 14:04:05.828344 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.512471 (* 0.3 = 0.153741 loss) | |
| I0409 14:04:05.828356 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.676874 | |
| I0409 14:04:05.828368 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.91591 | |
| I0409 14:04:05.828379 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.88203 | |
| I0409 14:04:05.828394 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.19853 (* 0.3 = 0.359559 loss) | |
| I0409 14:04:05.828408 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.312582 (* 0.3 = 0.0937746 loss) | |
| I0409 14:04:05.828420 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.805702 | |
| I0409 14:04:05.828433 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.951455 | |
| I0409 14:04:05.828444 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.901993 | |
| I0409 14:04:05.828457 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.850689 (* 1 = 0.850689 loss) | |
| I0409 14:04:05.828471 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.218278 (* 1 = 0.218278 loss) | |
| I0409 14:04:05.828497 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.512 | |
| I0409 14:04:05.828511 12249 solver.cpp:406] Test net output #16: total_confidence = 0.487499 | |
| I0409 14:04:06.201546 12249 solver.cpp:229] Iteration 160000, loss = 2.26377 | |
| I0409 14:04:06.201611 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.489796 | |
| I0409 14:04:06.201628 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 14:04:06.201642 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.714286 | |
| I0409 14:04:06.201658 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.86319 (* 0.3 = 0.558956 loss) | |
| I0409 14:04:06.201673 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.549064 (* 0.3 = 0.164719 loss) | |
| I0409 14:04:06.201686 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.530612 | |
| I0409 14:04:06.201699 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 14:04:06.201711 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.816327 | |
| I0409 14:04:06.201725 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.43007 (* 0.3 = 0.42902 loss) | |
| I0409 14:04:06.201740 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.421018 (* 0.3 = 0.126305 loss) | |
| I0409 14:04:06.201752 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.877551 | |
| I0409 14:04:06.201764 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 14:04:06.201776 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.938776 | |
| I0409 14:04:06.201791 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.435332 (* 1 = 0.435332 loss) | |
| I0409 14:04:06.201804 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.148128 (* 1 = 0.148128 loss) | |
| I0409 14:04:06.201817 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 14:04:06.201829 12249 solver.cpp:245] Train net output #16: total_confidence = 0.438017 | |
| I0409 14:04:06.201845 12249 sgd_solver.cpp:106] Iteration 160000, lr = 0.00771429 | |
| I0409 14:09:39.454536 12249 solver.cpp:229] Iteration 160500, loss = 2.2842 | |
| I0409 14:09:39.454751 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.346939 | |
| I0409 14:09:39.454773 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 14:09:39.454787 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.591837 | |
| I0409 14:09:39.454804 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.74855 (* 0.3 = 0.824564 loss) | |
| I0409 14:09:39.454819 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.82632 (* 0.3 = 0.247896 loss) | |
| I0409 14:09:39.454833 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.571429 | |
| I0409 14:09:39.454844 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 14:09:39.454857 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.77551 | |
| I0409 14:09:39.454871 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.85126 (* 0.3 = 0.555377 loss) | |
| I0409 14:09:39.454885 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.568187 (* 0.3 = 0.170456 loss) | |
| I0409 14:09:39.454897 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.734694 | |
| I0409 14:09:39.454910 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 14:09:39.454922 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.77551 | |
| I0409 14:09:39.454936 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.34626 (* 1 = 1.34626 loss) | |
| I0409 14:09:39.454951 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.394037 (* 1 = 0.394037 loss) | |
| I0409 14:09:39.454963 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 14:09:39.454975 12249 solver.cpp:245] Train net output #16: total_confidence = 0.396659 | |
| I0409 14:09:39.454991 12249 sgd_solver.cpp:106] Iteration 160500, lr = 0.00770714 | |
| I0409 14:15:12.795734 12249 solver.cpp:229] Iteration 161000, loss = 2.23153 | |
| I0409 14:15:12.796169 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.431818 | |
| I0409 14:15:12.796190 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 14:15:12.796205 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.659091 | |
| I0409 14:15:12.796222 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.05631 (* 0.3 = 0.616892 loss) | |
| I0409 14:15:12.796237 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.61104 (* 0.3 = 0.183312 loss) | |
| I0409 14:15:12.796250 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.704545 | |
| I0409 14:15:12.796262 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.914773 | |
| I0409 14:15:12.796274 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.909091 | |
| I0409 14:15:12.796289 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.0131 (* 0.3 = 0.30393 loss) | |
| I0409 14:15:12.796304 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.312648 (* 0.3 = 0.0937945 loss) | |
| I0409 14:15:12.796317 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.909091 | |
| I0409 14:15:12.796329 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 14:15:12.796342 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977273 | |
| I0409 14:15:12.796356 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.306128 (* 1 = 0.306128 loss) | |
| I0409 14:15:12.796371 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.080349 (* 1 = 0.080349 loss) | |
| I0409 14:15:12.796385 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 14:15:12.796396 12249 solver.cpp:245] Train net output #16: total_confidence = 0.630225 | |
| I0409 14:15:12.796412 12249 sgd_solver.cpp:106] Iteration 161000, lr = 0.0077 | |
| I0409 14:20:46.491402 12249 solver.cpp:229] Iteration 161500, loss = 2.31206 | |
| I0409 14:20:46.491564 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.307692 | |
| I0409 14:20:46.491585 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 14:20:46.491598 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.589744 | |
| I0409 14:20:46.491614 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.33681 (* 0.3 = 0.701043 loss) | |
| I0409 14:20:46.491631 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.727203 (* 0.3 = 0.218161 loss) | |
| I0409 14:20:46.491643 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.512821 | |
| I0409 14:20:46.491655 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 14:20:46.491668 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.794872 | |
| I0409 14:20:46.491683 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.78456 (* 0.3 = 0.535367 loss) | |
| I0409 14:20:46.491696 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.647888 (* 0.3 = 0.194366 loss) | |
| I0409 14:20:46.491709 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.641026 | |
| I0409 14:20:46.491721 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0409 14:20:46.491734 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.794872 | |
| I0409 14:20:46.491751 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.26517 (* 1 = 1.26517 loss) | |
| I0409 14:20:46.491766 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.384084 (* 1 = 0.384084 loss) | |
| I0409 14:20:46.491780 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 14:20:46.491791 12249 solver.cpp:245] Train net output #16: total_confidence = 0.287917 | |
| I0409 14:20:46.491806 12249 sgd_solver.cpp:106] Iteration 161500, lr = 0.00769286 | |
| I0409 14:26:20.221925 12249 solver.cpp:229] Iteration 162000, loss = 2.33649 | |
| I0409 14:26:20.222329 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.377358 | |
| I0409 14:26:20.222350 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 14:26:20.222364 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.679245 | |
| I0409 14:26:20.222381 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.00692 (* 0.3 = 0.602075 loss) | |
| I0409 14:26:20.222398 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.625021 (* 0.3 = 0.187506 loss) | |
| I0409 14:26:20.222411 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.509434 | |
| I0409 14:26:20.222424 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 14:26:20.222436 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.754717 | |
| I0409 14:26:20.222450 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.65807 (* 0.3 = 0.497421 loss) | |
| I0409 14:26:20.222465 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.516795 (* 0.3 = 0.155039 loss) | |
| I0409 14:26:20.222476 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.811321 | |
| I0409 14:26:20.222489 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 14:26:20.222501 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.962264 | |
| I0409 14:26:20.222517 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.610604 (* 1 = 0.610604 loss) | |
| I0409 14:26:20.222530 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.188794 (* 1 = 0.188794 loss) | |
| I0409 14:26:20.222543 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 14:26:20.222559 12249 solver.cpp:245] Train net output #16: total_confidence = 0.324072 | |
| I0409 14:26:20.222587 12249 sgd_solver.cpp:106] Iteration 162000, lr = 0.00768571 | |
| I0409 14:31:54.058820 12249 solver.cpp:229] Iteration 162500, loss = 2.27149 | |
| I0409 14:31:54.059128 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.465116 | |
| I0409 14:31:54.059147 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 14:31:54.059162 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.72093 | |
| I0409 14:31:54.059180 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.66474 (* 0.3 = 0.499421 loss) | |
| I0409 14:31:54.059195 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.477457 (* 0.3 = 0.143237 loss) | |
| I0409 14:31:54.059206 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55814 | |
| I0409 14:31:54.059219 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 14:31:54.059231 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.860465 | |
| I0409 14:31:54.059247 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.30139 (* 0.3 = 0.390418 loss) | |
| I0409 14:31:54.059262 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.357512 (* 0.3 = 0.107254 loss) | |
| I0409 14:31:54.059278 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.837209 | |
| I0409 14:31:54.059291 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 14:31:54.059303 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.976744 | |
| I0409 14:31:54.059319 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.558073 (* 1 = 0.558073 loss) | |
| I0409 14:31:54.059332 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.173829 (* 1 = 0.173829 loss) | |
| I0409 14:31:54.059345 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 14:31:54.059357 12249 solver.cpp:245] Train net output #16: total_confidence = 0.293307 | |
| I0409 14:31:54.059372 12249 sgd_solver.cpp:106] Iteration 162500, lr = 0.00767857 | |
| I0409 14:37:28.431324 12249 solver.cpp:229] Iteration 163000, loss = 2.23081 | |
| I0409 14:37:28.431529 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.386364 | |
| I0409 14:37:28.431548 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 14:37:28.431561 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.75 | |
| I0409 14:37:28.431578 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.69933 (* 0.3 = 0.509799 loss) | |
| I0409 14:37:28.431593 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.45658 (* 0.3 = 0.136974 loss) | |
| I0409 14:37:28.431607 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.681818 | |
| I0409 14:37:28.431619 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.914773 | |
| I0409 14:37:28.431632 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.909091 | |
| I0409 14:37:28.431645 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.07328 (* 0.3 = 0.321983 loss) | |
| I0409 14:37:28.431660 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.30553 (* 0.3 = 0.0916589 loss) | |
| I0409 14:37:28.431674 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.977273 | |
| I0409 14:37:28.431685 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.994318 | |
| I0409 14:37:28.431697 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.977273 | |
| I0409 14:37:28.431712 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.206694 (* 1 = 0.206694 loss) | |
| I0409 14:37:28.431726 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0678437 (* 1 = 0.0678437 loss) | |
| I0409 14:37:28.431740 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 14:37:28.431756 12249 solver.cpp:245] Train net output #16: total_confidence = 0.51224 | |
| I0409 14:37:28.431771 12249 sgd_solver.cpp:106] Iteration 163000, lr = 0.00767143 | |
| I0409 14:43:02.733757 12249 solver.cpp:229] Iteration 163500, loss = 2.29856 | |
| I0409 14:43:02.734076 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.226415 | |
| I0409 14:43:02.734096 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.767045 | |
| I0409 14:43:02.734109 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.45283 | |
| I0409 14:43:02.734127 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.60492 (* 0.3 = 0.781477 loss) | |
| I0409 14:43:02.734141 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.804968 (* 0.3 = 0.24149 loss) | |
| I0409 14:43:02.734154 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.566038 | |
| I0409 14:43:02.734166 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 14:43:02.734179 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.830189 | |
| I0409 14:43:02.734194 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.58735 (* 0.3 = 0.476205 loss) | |
| I0409 14:43:02.734207 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.494376 (* 0.3 = 0.148313 loss) | |
| I0409 14:43:02.734220 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.849057 | |
| I0409 14:43:02.734231 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 14:43:02.734243 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.962264 | |
| I0409 14:43:02.734258 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.510207 (* 1 = 0.510207 loss) | |
| I0409 14:43:02.734272 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.163189 (* 1 = 0.163189 loss) | |
| I0409 14:43:02.734287 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 14:43:02.734302 12249 solver.cpp:245] Train net output #16: total_confidence = 0.328236 | |
| I0409 14:43:02.734316 12249 sgd_solver.cpp:106] Iteration 163500, lr = 0.00766429 | |
| I0409 14:48:37.201575 12249 solver.cpp:229] Iteration 164000, loss = 2.27024 | |
| I0409 14:48:37.201735 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 14:48:37.201756 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 14:48:37.201778 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.586957 | |
| I0409 14:48:37.201807 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.98586 (* 0.3 = 0.595757 loss) | |
| I0409 14:48:37.201836 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.600286 (* 0.3 = 0.180086 loss) | |
| I0409 14:48:37.201859 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.608696 | |
| I0409 14:48:37.201885 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 14:48:37.201910 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.804348 | |
| I0409 14:48:37.201936 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.43329 (* 0.3 = 0.429987 loss) | |
| I0409 14:48:37.201963 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.437193 (* 0.3 = 0.131158 loss) | |
| I0409 14:48:37.201987 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.804348 | |
| I0409 14:48:37.202010 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 14:48:37.202031 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 14:48:37.202057 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.574554 (* 1 = 0.574554 loss) | |
| I0409 14:48:37.202083 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.156374 (* 1 = 0.156374 loss) | |
| I0409 14:48:37.202105 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 14:48:37.202126 12249 solver.cpp:245] Train net output #16: total_confidence = 0.571316 | |
| I0409 14:48:37.202150 12249 sgd_solver.cpp:106] Iteration 164000, lr = 0.00765714 | |
| I0409 14:54:11.695454 12249 solver.cpp:229] Iteration 164500, loss = 2.34068 | |
| I0409 14:54:11.695780 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 14:54:11.695801 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 14:54:11.695814 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.76087 | |
| I0409 14:54:11.695832 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.89022 (* 0.3 = 0.567065 loss) | |
| I0409 14:54:11.695847 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.570409 (* 0.3 = 0.171123 loss) | |
| I0409 14:54:11.695860 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.652174 | |
| I0409 14:54:11.695873 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 14:54:11.695886 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.869565 | |
| I0409 14:54:11.695901 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.27737 (* 0.3 = 0.38321 loss) | |
| I0409 14:54:11.695915 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.398239 (* 0.3 = 0.119472 loss) | |
| I0409 14:54:11.695929 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.869565 | |
| I0409 14:54:11.695941 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 14:54:11.695953 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 14:54:11.695968 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.680274 (* 1 = 0.680274 loss) | |
| I0409 14:54:11.695982 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.196681 (* 1 = 0.196681 loss) | |
| I0409 14:54:11.695996 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 14:54:11.696007 12249 solver.cpp:245] Train net output #16: total_confidence = 0.487209 | |
| I0409 14:54:11.696028 12249 sgd_solver.cpp:106] Iteration 164500, lr = 0.00765 | |
| I0409 14:59:45.760294 12249 solver.cpp:338] Iteration 165000, Testing net (#0) | |
| I0409 15:00:27.650741 12249 solver.cpp:393] Test loss: 1.99828 | |
| I0409 15:00:27.650887 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.482655 | |
| I0409 15:00:27.650907 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.865139 | |
| I0409 15:00:27.650920 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.773492 | |
| I0409 15:00:27.650938 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.71535 (* 0.3 = 0.514605 loss) | |
| I0409 15:00:27.650952 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.457189 (* 0.3 = 0.137157 loss) | |
| I0409 15:00:27.650964 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.706928 | |
| I0409 15:00:27.650976 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.917548 | |
| I0409 15:00:27.650987 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.888044 | |
| I0409 15:00:27.651001 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.07431 (* 0.3 = 0.322294 loss) | |
| I0409 15:00:27.651015 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.297585 (* 0.3 = 0.0892755 loss) | |
| I0409 15:00:27.651027 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.824106 | |
| I0409 15:00:27.651039 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.953409 | |
| I0409 15:00:27.651051 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.912902 | |
| I0409 15:00:27.651064 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.737889 (* 1 = 0.737889 loss) | |
| I0409 15:00:27.651078 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.197062 (* 1 = 0.197062 loss) | |
| I0409 15:00:27.651090 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.543 | |
| I0409 15:00:27.651103 12249 solver.cpp:406] Test net output #16: total_confidence = 0.450222 | |
| I0409 15:00:28.029608 12249 solver.cpp:229] Iteration 165000, loss = 2.25822 | |
| I0409 15:00:28.029690 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.5 | |
| I0409 15:00:28.029708 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.869318 | |
| I0409 15:00:28.029722 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.76087 | |
| I0409 15:00:28.029742 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.62685 (* 0.3 = 0.488054 loss) | |
| I0409 15:00:28.029757 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.464126 (* 0.3 = 0.139238 loss) | |
| I0409 15:00:28.029769 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.695652 | |
| I0409 15:00:28.029783 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.920455 | |
| I0409 15:00:28.029794 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.934783 | |
| I0409 15:00:28.029808 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.0398 (* 0.3 = 0.31194 loss) | |
| I0409 15:00:28.029824 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.290173 (* 0.3 = 0.0870518 loss) | |
| I0409 15:00:28.029836 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.913043 | |
| I0409 15:00:28.029850 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 15:00:28.029861 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 15:00:28.029875 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.341514 (* 1 = 0.341514 loss) | |
| I0409 15:00:28.029891 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0911226 (* 1 = 0.0911226 loss) | |
| I0409 15:00:28.029903 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 15:00:28.029916 12249 solver.cpp:245] Train net output #16: total_confidence = 0.66853 | |
| I0409 15:00:28.029932 12249 sgd_solver.cpp:106] Iteration 165000, lr = 0.00764286 | |
| I0409 15:06:01.999687 12249 solver.cpp:229] Iteration 165500, loss = 2.24749 | |
| I0409 15:06:02.000077 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.555556 | |
| I0409 15:06:02.000098 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.869318 | |
| I0409 15:06:02.000113 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.8 | |
| I0409 15:06:02.000129 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.54356 (* 0.3 = 0.463068 loss) | |
| I0409 15:06:02.000144 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.445944 (* 0.3 = 0.133783 loss) | |
| I0409 15:06:02.000157 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.755556 | |
| I0409 15:06:02.000169 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.914773 | |
| I0409 15:06:02.000182 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.933333 | |
| I0409 15:06:02.000196 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.783561 (* 0.3 = 0.235068 loss) | |
| I0409 15:06:02.000211 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.236266 (* 0.3 = 0.0708799 loss) | |
| I0409 15:06:02.000224 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.888889 | |
| I0409 15:06:02.000236 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 15:06:02.000248 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 15:06:02.000263 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.265616 (* 1 = 0.265616 loss) | |
| I0409 15:06:02.000278 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0925275 (* 1 = 0.0925275 loss) | |
| I0409 15:06:02.000290 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 15:06:02.000303 12249 solver.cpp:245] Train net output #16: total_confidence = 0.559605 | |
| I0409 15:06:02.000319 12249 sgd_solver.cpp:106] Iteration 165500, lr = 0.00763571 | |
| I0409 15:11:35.574129 12249 solver.cpp:229] Iteration 166000, loss = 2.3048 | |
| I0409 15:11:35.574429 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.473684 | |
| I0409 15:11:35.574450 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 15:11:35.574463 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.578947 | |
| I0409 15:11:35.574481 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.12211 (* 0.3 = 0.636634 loss) | |
| I0409 15:11:35.574496 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.569759 (* 0.3 = 0.170928 loss) | |
| I0409 15:11:35.574508 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.342105 | |
| I0409 15:11:35.574522 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.823864 | |
| I0409 15:11:35.574532 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.605263 | |
| I0409 15:11:35.574547 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.06108 (* 0.3 = 0.618325 loss) | |
| I0409 15:11:35.574560 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.59102 (* 0.3 = 0.177306 loss) | |
| I0409 15:11:35.574573 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.763158 | |
| I0409 15:11:35.574585 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 15:11:35.574597 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.894737 | |
| I0409 15:11:35.574612 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.961839 (* 1 = 0.961839 loss) | |
| I0409 15:11:35.574626 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.232704 (* 1 = 0.232704 loss) | |
| I0409 15:11:35.574640 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.125 | |
| I0409 15:11:35.574652 12249 solver.cpp:245] Train net output #16: total_confidence = 0.264307 | |
| I0409 15:11:35.574667 12249 sgd_solver.cpp:106] Iteration 166000, lr = 0.00762857 | |
| I0409 15:17:08.923501 12249 solver.cpp:229] Iteration 166500, loss = 2.25597 | |
| I0409 15:17:08.923722 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.372093 | |
| I0409 15:17:08.923746 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 15:17:08.923761 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.674419 | |
| I0409 15:17:08.923779 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.87631 (* 0.3 = 0.562893 loss) | |
| I0409 15:17:08.923794 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.530137 (* 0.3 = 0.159041 loss) | |
| I0409 15:17:08.923807 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.627907 | |
| I0409 15:17:08.923820 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 15:17:08.923832 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.883721 | |
| I0409 15:17:08.923846 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.10344 (* 0.3 = 0.331031 loss) | |
| I0409 15:17:08.923861 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.363265 (* 0.3 = 0.10898 loss) | |
| I0409 15:17:08.923873 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.860465 | |
| I0409 15:17:08.923885 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 15:17:08.923897 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 15:17:08.923913 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.374674 (* 1 = 0.374674 loss) | |
| I0409 15:17:08.923928 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.102284 (* 1 = 0.102284 loss) | |
| I0409 15:17:08.923940 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 15:17:08.923952 12249 solver.cpp:245] Train net output #16: total_confidence = 0.46858 | |
| I0409 15:17:08.923967 12249 sgd_solver.cpp:106] Iteration 166500, lr = 0.00762143 | |
| I0409 15:22:42.325626 12249 solver.cpp:229] Iteration 167000, loss = 2.28188 | |
| I0409 15:22:42.325924 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.521739 | |
| I0409 15:22:42.325944 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.863636 | |
| I0409 15:22:42.325958 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.652174 | |
| I0409 15:22:42.325973 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.6798 (* 0.3 = 0.503941 loss) | |
| I0409 15:22:42.325989 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.47638 (* 0.3 = 0.142914 loss) | |
| I0409 15:22:42.326001 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.652174 | |
| I0409 15:22:42.326014 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 15:22:42.326026 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.891304 | |
| I0409 15:22:42.326040 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.19718 (* 0.3 = 0.359153 loss) | |
| I0409 15:22:42.326058 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.32813 (* 0.3 = 0.098439 loss) | |
| I0409 15:22:42.326071 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826087 | |
| I0409 15:22:42.326084 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 15:22:42.326097 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 15:22:42.326110 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.546257 (* 1 = 0.546257 loss) | |
| I0409 15:22:42.326124 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.164871 (* 1 = 0.164871 loss) | |
| I0409 15:22:42.326138 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 15:22:42.326150 12249 solver.cpp:245] Train net output #16: total_confidence = 0.483983 | |
| I0409 15:22:42.326165 12249 sgd_solver.cpp:106] Iteration 167000, lr = 0.00761429 | |
| I0409 15:28:15.688665 12249 solver.cpp:229] Iteration 167500, loss = 2.3039 | |
| I0409 15:28:15.688873 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.23913 | |
| I0409 15:28:15.688894 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 15:28:15.688907 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.413043 | |
| I0409 15:28:15.688925 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.40785 (* 0.3 = 0.722356 loss) | |
| I0409 15:28:15.688941 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.65607 (* 0.3 = 0.196821 loss) | |
| I0409 15:28:15.688953 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.413043 | |
| I0409 15:28:15.688966 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 15:28:15.688978 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.673913 | |
| I0409 15:28:15.688992 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.20008 (* 0.3 = 0.660025 loss) | |
| I0409 15:28:15.689007 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.597657 (* 0.3 = 0.179297 loss) | |
| I0409 15:28:15.689019 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.652174 | |
| I0409 15:28:15.689031 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 15:28:15.689043 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.717391 | |
| I0409 15:28:15.689059 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.52246 (* 1 = 1.52246 loss) | |
| I0409 15:28:15.689072 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.420194 (* 1 = 0.420194 loss) | |
| I0409 15:28:15.689085 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 15:28:15.689097 12249 solver.cpp:245] Train net output #16: total_confidence = 0.281178 | |
| I0409 15:28:15.689112 12249 sgd_solver.cpp:106] Iteration 167500, lr = 0.00760714 | |
| I0409 15:33:49.066294 12249 solver.cpp:229] Iteration 168000, loss = 2.27522 | |
| I0409 15:33:49.066591 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.5 | |
| I0409 15:33:49.066612 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 15:33:49.066627 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.65 | |
| I0409 15:33:49.066642 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.93872 (* 0.3 = 0.581615 loss) | |
| I0409 15:33:49.066658 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.604402 (* 0.3 = 0.18132 loss) | |
| I0409 15:33:49.066670 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.625 | |
| I0409 15:33:49.066682 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 15:33:49.066694 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.925 | |
| I0409 15:33:49.066709 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.977455 (* 0.3 = 0.293236 loss) | |
| I0409 15:33:49.066722 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.299034 (* 0.3 = 0.0897101 loss) | |
| I0409 15:33:49.066735 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.9 | |
| I0409 15:33:49.066747 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 15:33:49.066759 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.95 | |
| I0409 15:33:49.066773 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.464056 (* 1 = 0.464056 loss) | |
| I0409 15:33:49.066788 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.152186 (* 1 = 0.152186 loss) | |
| I0409 15:33:49.066800 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 15:33:49.066812 12249 solver.cpp:245] Train net output #16: total_confidence = 0.440329 | |
| I0409 15:33:49.066826 12249 sgd_solver.cpp:106] Iteration 168000, lr = 0.0076 | |
| I0409 15:39:22.425823 12249 solver.cpp:229] Iteration 168500, loss = 2.27551 | |
| I0409 15:39:22.425983 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.382979 | |
| I0409 15:39:22.426014 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 15:39:22.426041 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.744681 | |
| I0409 15:39:22.426065 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.9121 (* 0.3 = 0.573631 loss) | |
| I0409 15:39:22.426081 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.721874 (* 0.3 = 0.216562 loss) | |
| I0409 15:39:22.426095 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.531915 | |
| I0409 15:39:22.426106 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 15:39:22.426120 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.829787 | |
| I0409 15:39:22.426133 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.45706 (* 0.3 = 0.437117 loss) | |
| I0409 15:39:22.426147 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.614846 (* 0.3 = 0.184454 loss) | |
| I0409 15:39:22.426161 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.893617 | |
| I0409 15:39:22.426172 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 15:39:22.426184 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0409 15:39:22.426198 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.507466 (* 1 = 0.507466 loss) | |
| I0409 15:39:22.426213 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.386532 (* 1 = 0.386532 loss) | |
| I0409 15:39:22.426224 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 15:39:22.426236 12249 solver.cpp:245] Train net output #16: total_confidence = 0.325402 | |
| I0409 15:39:22.426251 12249 sgd_solver.cpp:106] Iteration 168500, lr = 0.00759286 | |
| I0409 15:44:55.788681 12249 solver.cpp:229] Iteration 169000, loss = 2.29205 | |
| I0409 15:44:55.788965 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.425532 | |
| I0409 15:44:55.788995 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 15:44:55.789019 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.595745 | |
| I0409 15:44:55.789048 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95132 (* 0.3 = 0.585397 loss) | |
| I0409 15:44:55.789075 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.580624 (* 0.3 = 0.174187 loss) | |
| I0409 15:44:55.789101 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.489362 | |
| I0409 15:44:55.789125 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 15:44:55.789147 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.829787 | |
| I0409 15:44:55.789172 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.59336 (* 0.3 = 0.478009 loss) | |
| I0409 15:44:55.789196 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.515803 (* 0.3 = 0.154741 loss) | |
| I0409 15:44:55.789218 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.702128 | |
| I0409 15:44:55.789240 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 15:44:55.789263 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.851064 | |
| I0409 15:44:55.789289 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.962786 (* 1 = 0.962786 loss) | |
| I0409 15:44:55.789314 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.288002 (* 1 = 0.288002 loss) | |
| I0409 15:44:55.789335 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 15:44:55.789357 12249 solver.cpp:245] Train net output #16: total_confidence = 0.288824 | |
| I0409 15:44:55.789383 12249 sgd_solver.cpp:106] Iteration 169000, lr = 0.00758571 | |
| I0409 15:50:29.163164 12249 solver.cpp:229] Iteration 169500, loss = 2.30099 | |
| I0409 15:50:29.163352 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.302326 | |
| I0409 15:50:29.163381 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 15:50:29.163403 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.55814 | |
| I0409 15:50:29.163431 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.12753 (* 0.3 = 0.638258 loss) | |
| I0409 15:50:29.163460 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.668909 (* 0.3 = 0.200673 loss) | |
| I0409 15:50:29.163481 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.604651 | |
| I0409 15:50:29.163503 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 15:50:29.163524 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.813953 | |
| I0409 15:50:29.163550 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.34458 (* 0.3 = 0.403374 loss) | |
| I0409 15:50:29.163578 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.444045 (* 0.3 = 0.133213 loss) | |
| I0409 15:50:29.163599 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.883721 | |
| I0409 15:50:29.163620 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 15:50:29.163640 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0409 15:50:29.163666 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.501737 (* 1 = 0.501737 loss) | |
| I0409 15:50:29.163691 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.163493 (* 1 = 0.163493 loss) | |
| I0409 15:50:29.163712 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 15:50:29.163734 12249 solver.cpp:245] Train net output #16: total_confidence = 0.484239 | |
| I0409 15:50:29.163763 12249 sgd_solver.cpp:106] Iteration 169500, lr = 0.00757857 | |
| I0409 15:56:02.146132 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_170000.caffemodel | |
| I0409 15:56:02.644613 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_170000.solverstate | |
| I0409 15:56:02.891125 12249 solver.cpp:338] Iteration 170000, Testing net (#0) | |
| I0409 15:56:43.888973 12249 solver.cpp:393] Test loss: 2.00803 | |
| I0409 15:56:43.889084 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.491253 | |
| I0409 15:56:43.889103 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.864639 | |
| I0409 15:56:43.889117 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.769775 | |
| I0409 15:56:43.889132 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.76791 (* 0.3 = 0.530373 loss) | |
| I0409 15:56:43.889147 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.476219 (* 0.3 = 0.142866 loss) | |
| I0409 15:56:43.889159 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.724651 | |
| I0409 15:56:43.889171 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.917865 | |
| I0409 15:56:43.889183 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.891795 | |
| I0409 15:56:43.889195 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.05954 (* 0.3 = 0.317862 loss) | |
| I0409 15:56:43.889210 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.306369 (* 0.3 = 0.0919106 loss) | |
| I0409 15:56:43.889221 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.827024 | |
| I0409 15:56:43.889233 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.953455 | |
| I0409 15:56:43.889245 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.912772 | |
| I0409 15:56:43.889258 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.729098 (* 1 = 0.729098 loss) | |
| I0409 15:56:43.889272 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.195917 (* 1 = 0.195917 loss) | |
| I0409 15:56:43.889284 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.557 | |
| I0409 15:56:43.889295 12249 solver.cpp:406] Test net output #16: total_confidence = 0.493723 | |
| I0409 15:56:44.262045 12249 solver.cpp:229] Iteration 170000, loss = 2.32927 | |
| I0409 15:56:44.262104 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.377358 | |
| I0409 15:56:44.262121 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 15:56:44.262135 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.566038 | |
| I0409 15:56:44.262151 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.37741 (* 0.3 = 0.713224 loss) | |
| I0409 15:56:44.262166 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.755261 (* 0.3 = 0.226578 loss) | |
| I0409 15:56:44.262178 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.396226 | |
| I0409 15:56:44.262190 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.8125 | |
| I0409 15:56:44.262202 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.641509 | |
| I0409 15:56:44.262217 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.93761 (* 0.3 = 0.581284 loss) | |
| I0409 15:56:44.262231 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.609375 (* 0.3 = 0.182813 loss) | |
| I0409 15:56:44.262246 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.641509 | |
| I0409 15:56:44.262260 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0409 15:56:44.262272 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.811321 | |
| I0409 15:56:44.262286 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.15219 (* 1 = 1.15219 loss) | |
| I0409 15:56:44.262300 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.368796 (* 1 = 0.368796 loss) | |
| I0409 15:56:44.262313 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 15:56:44.262326 12249 solver.cpp:245] Train net output #16: total_confidence = 0.186865 | |
| I0409 15:56:44.262341 12249 sgd_solver.cpp:106] Iteration 170000, lr = 0.00757143 | |
| I0409 16:02:17.567322 12249 solver.cpp:229] Iteration 170500, loss = 2.27431 | |
| I0409 16:02:17.567651 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.416667 | |
| I0409 16:02:17.567672 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 16:02:17.567685 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.666667 | |
| I0409 16:02:17.567703 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95287 (* 0.3 = 0.585862 loss) | |
| I0409 16:02:17.567718 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.598231 (* 0.3 = 0.179469 loss) | |
| I0409 16:02:17.567731 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.625 | |
| I0409 16:02:17.567747 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 16:02:17.567759 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.854167 | |
| I0409 16:02:17.567773 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.16082 (* 0.3 = 0.348246 loss) | |
| I0409 16:02:17.567788 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.377697 (* 0.3 = 0.113309 loss) | |
| I0409 16:02:17.567801 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.916667 | |
| I0409 16:02:17.567813 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 16:02:17.567826 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 16:02:17.567839 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.32778 (* 1 = 0.32778 loss) | |
| I0409 16:02:17.567853 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.105041 (* 1 = 0.105041 loss) | |
| I0409 16:02:17.567865 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 16:02:17.567878 12249 solver.cpp:245] Train net output #16: total_confidence = 0.398262 | |
| I0409 16:02:17.567893 12249 sgd_solver.cpp:106] Iteration 170500, lr = 0.00756429 | |
| I0409 16:07:50.954659 12249 solver.cpp:229] Iteration 171000, loss = 2.25821 | |
| I0409 16:07:50.954790 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.38 | |
| I0409 16:07:50.954810 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 16:07:50.954823 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.62 | |
| I0409 16:07:50.954840 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.95873 (* 0.3 = 0.587618 loss) | |
| I0409 16:07:50.954855 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.631927 (* 0.3 = 0.189578 loss) | |
| I0409 16:07:50.954869 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.7 | |
| I0409 16:07:50.954880 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 16:07:50.954892 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.9 | |
| I0409 16:07:50.954906 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.06739 (* 0.3 = 0.320216 loss) | |
| I0409 16:07:50.954921 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.329505 (* 0.3 = 0.0988516 loss) | |
| I0409 16:07:50.954933 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.94 | |
| I0409 16:07:50.954946 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 16:07:50.954957 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 16:07:50.954972 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.284972 (* 1 = 0.284972 loss) | |
| I0409 16:07:50.954987 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.109406 (* 1 = 0.109406 loss) | |
| I0409 16:07:50.954999 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 16:07:50.955011 12249 solver.cpp:245] Train net output #16: total_confidence = 0.362988 | |
| I0409 16:07:50.955025 12249 sgd_solver.cpp:106] Iteration 171000, lr = 0.00755714 | |
| I0409 16:13:24.319866 12249 solver.cpp:229] Iteration 171500, loss = 2.24767 | |
| I0409 16:13:24.320251 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.3 | |
| I0409 16:13:24.320273 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 16:13:24.320287 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.425 | |
| I0409 16:13:24.320304 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.48585 (* 0.3 = 0.745756 loss) | |
| I0409 16:13:24.320319 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.680268 (* 0.3 = 0.20408 loss) | |
| I0409 16:13:24.320333 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.6 | |
| I0409 16:13:24.320344 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 16:13:24.320356 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.825 | |
| I0409 16:13:24.320370 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.41377 (* 0.3 = 0.424131 loss) | |
| I0409 16:13:24.320386 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.404625 (* 0.3 = 0.121387 loss) | |
| I0409 16:13:24.320399 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.9 | |
| I0409 16:13:24.320411 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 16:13:24.320423 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.975 | |
| I0409 16:13:24.320437 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.489545 (* 1 = 0.489545 loss) | |
| I0409 16:13:24.320452 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.115556 (* 1 = 0.115556 loss) | |
| I0409 16:13:24.320466 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 16:13:24.320477 12249 solver.cpp:245] Train net output #16: total_confidence = 0.460575 | |
| I0409 16:13:24.320511 12249 sgd_solver.cpp:106] Iteration 171500, lr = 0.00755 | |
| I0409 16:18:57.687446 12249 solver.cpp:229] Iteration 172000, loss = 2.25689 | |
| I0409 16:18:57.687577 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.268293 | |
| I0409 16:18:57.687607 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 16:18:57.687629 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.560976 | |
| I0409 16:18:57.687657 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.53152 (* 0.3 = 0.759455 loss) | |
| I0409 16:18:57.687685 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.745537 (* 0.3 = 0.223661 loss) | |
| I0409 16:18:57.687706 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.560976 | |
| I0409 16:18:57.687728 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 16:18:57.687752 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.804878 | |
| I0409 16:18:57.687778 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.52469 (* 0.3 = 0.457408 loss) | |
| I0409 16:18:57.687805 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.493221 (* 0.3 = 0.147966 loss) | |
| I0409 16:18:57.687827 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.829268 | |
| I0409 16:18:57.687849 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 16:18:57.687868 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.853659 | |
| I0409 16:18:57.687893 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.10772 (* 1 = 1.10772 loss) | |
| I0409 16:18:57.687918 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.332266 (* 1 = 0.332266 loss) | |
| I0409 16:18:57.687939 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 16:18:57.687963 12249 solver.cpp:245] Train net output #16: total_confidence = 0.50855 | |
| I0409 16:18:57.687988 12249 sgd_solver.cpp:106] Iteration 172000, lr = 0.00754286 | |
| I0409 16:24:31.069054 12249 solver.cpp:229] Iteration 172500, loss = 2.25641 | |
| I0409 16:24:31.069413 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.479167 | |
| I0409 16:24:31.069442 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 16:24:31.069465 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.6875 | |
| I0409 16:24:31.069492 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.97574 (* 0.3 = 0.592723 loss) | |
| I0409 16:24:31.069522 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.57525 (* 0.3 = 0.172575 loss) | |
| I0409 16:24:31.069543 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.625 | |
| I0409 16:24:31.069564 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 16:24:31.069584 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.770833 | |
| I0409 16:24:31.069609 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.29256 (* 0.3 = 0.387768 loss) | |
| I0409 16:24:31.069631 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.381139 (* 0.3 = 0.114342 loss) | |
| I0409 16:24:31.069653 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.854167 | |
| I0409 16:24:31.069676 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 16:24:31.069699 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.916667 | |
| I0409 16:24:31.069722 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.665848 (* 1 = 0.665848 loss) | |
| I0409 16:24:31.069752 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.187122 (* 1 = 0.187122 loss) | |
| I0409 16:24:31.069777 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 16:24:31.069799 12249 solver.cpp:245] Train net output #16: total_confidence = 0.521498 | |
| I0409 16:24:31.069823 12249 sgd_solver.cpp:106] Iteration 172500, lr = 0.00753571 | |
| I0409 16:30:04.460060 12249 solver.cpp:229] Iteration 173000, loss = 2.20033 | |
| I0409 16:30:04.460201 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.265306 | |
| I0409 16:30:04.460221 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.772727 | |
| I0409 16:30:04.460233 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.571429 | |
| I0409 16:30:04.460250 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.58375 (* 0.3 = 0.775125 loss) | |
| I0409 16:30:04.460265 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.787128 (* 0.3 = 0.236138 loss) | |
| I0409 16:30:04.460278 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.408163 | |
| I0409 16:30:04.460290 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 16:30:04.460302 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.77551 | |
| I0409 16:30:04.460316 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.77528 (* 0.3 = 0.532585 loss) | |
| I0409 16:30:04.460330 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.546288 (* 0.3 = 0.163886 loss) | |
| I0409 16:30:04.460343 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.714286 | |
| I0409 16:30:04.460355 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 16:30:04.460367 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.938776 | |
| I0409 16:30:04.460382 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.843843 (* 1 = 0.843843 loss) | |
| I0409 16:30:04.460397 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.298738 (* 1 = 0.298738 loss) | |
| I0409 16:30:04.460408 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 16:30:04.460420 12249 solver.cpp:245] Train net output #16: total_confidence = 0.37599 | |
| I0409 16:30:04.460436 12249 sgd_solver.cpp:106] Iteration 173000, lr = 0.00752857 | |
| I0409 16:35:37.798632 12249 solver.cpp:229] Iteration 173500, loss = 2.27872 | |
| I0409 16:35:37.799800 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.358974 | |
| I0409 16:35:37.799824 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 16:35:37.799839 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.564103 | |
| I0409 16:35:37.799855 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.42109 (* 0.3 = 0.726328 loss) | |
| I0409 16:35:37.799871 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.666982 (* 0.3 = 0.200095 loss) | |
| I0409 16:35:37.799885 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.538462 | |
| I0409 16:35:37.799897 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 16:35:37.799909 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.717949 | |
| I0409 16:35:37.799924 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.82517 (* 0.3 = 0.547551 loss) | |
| I0409 16:35:37.799939 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.661493 (* 0.3 = 0.198448 loss) | |
| I0409 16:35:37.799952 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.641026 | |
| I0409 16:35:37.799964 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0409 16:35:37.799978 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.769231 | |
| I0409 16:35:37.799991 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.44139 (* 1 = 1.44139 loss) | |
| I0409 16:35:37.800005 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.459048 (* 1 = 0.459048 loss) | |
| I0409 16:35:37.800019 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 16:35:37.800031 12249 solver.cpp:245] Train net output #16: total_confidence = 0.408684 | |
| I0409 16:35:37.800046 12249 sgd_solver.cpp:106] Iteration 173500, lr = 0.00752143 | |
| I0409 16:41:11.176165 12249 solver.cpp:229] Iteration 174000, loss = 2.27186 | |
| I0409 16:41:11.176314 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.384615 | |
| I0409 16:41:11.176336 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 16:41:11.176348 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.692308 | |
| I0409 16:41:11.176365 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.06781 (* 0.3 = 0.620343 loss) | |
| I0409 16:41:11.176381 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.557763 (* 0.3 = 0.167329 loss) | |
| I0409 16:41:11.176394 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.564103 | |
| I0409 16:41:11.176406 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 16:41:11.176419 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.794872 | |
| I0409 16:41:11.176434 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.38643 (* 0.3 = 0.415928 loss) | |
| I0409 16:41:11.176448 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.444219 (* 0.3 = 0.133266 loss) | |
| I0409 16:41:11.176461 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.846154 | |
| I0409 16:41:11.176473 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 16:41:11.176499 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.948718 | |
| I0409 16:41:11.176517 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.566763 (* 1 = 0.566763 loss) | |
| I0409 16:41:11.176530 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.194002 (* 1 = 0.194002 loss) | |
| I0409 16:41:11.176543 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 16:41:11.176555 12249 solver.cpp:245] Train net output #16: total_confidence = 0.30383 | |
| I0409 16:41:11.176570 12249 sgd_solver.cpp:106] Iteration 174000, lr = 0.00751429 | |
| I0409 16:46:44.559134 12249 solver.cpp:229] Iteration 174500, loss = 2.23068 | |
| I0409 16:46:44.559474 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.27027 | |
| I0409 16:46:44.559495 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 16:46:44.559509 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.675676 | |
| I0409 16:46:44.559525 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.21491 (* 0.3 = 0.664473 loss) | |
| I0409 16:46:44.559541 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.602088 (* 0.3 = 0.180626 loss) | |
| I0409 16:46:44.559553 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.540541 | |
| I0409 16:46:44.559566 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 16:46:44.559578 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.756757 | |
| I0409 16:46:44.559592 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.44097 (* 0.3 = 0.43229 loss) | |
| I0409 16:46:44.559607 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.468682 (* 0.3 = 0.140605 loss) | |
| I0409 16:46:44.559620 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.837838 | |
| I0409 16:46:44.559633 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 16:46:44.559645 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.918919 | |
| I0409 16:46:44.559659 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.7427 (* 1 = 0.7427 loss) | |
| I0409 16:46:44.559674 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.203948 (* 1 = 0.203948 loss) | |
| I0409 16:46:44.559686 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 16:46:44.559698 12249 solver.cpp:245] Train net output #16: total_confidence = 0.364475 | |
| I0409 16:46:44.559713 12249 sgd_solver.cpp:106] Iteration 174500, lr = 0.00750714 | |
| I0409 16:52:17.875975 12249 solver.cpp:338] Iteration 175000, Testing net (#0) | |
| I0409 16:52:59.158888 12249 solver.cpp:393] Test loss: 1.9786 | |
| I0409 16:52:59.159024 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.520521 | |
| I0409 16:52:59.159044 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.870594 | |
| I0409 16:52:59.159057 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.785487 | |
| I0409 16:52:59.159075 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.66403 (* 0.3 = 0.49921 loss) | |
| I0409 16:52:59.159090 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.452655 (* 0.3 = 0.135797 loss) | |
| I0409 16:52:59.159101 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.681281 | |
| I0409 16:52:59.159113 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.918183 | |
| I0409 16:52:59.159124 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.890832 | |
| I0409 16:52:59.159138 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.12115 (* 0.3 = 0.336346 loss) | |
| I0409 16:52:59.159152 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.292197 (* 0.3 = 0.0876592 loss) | |
| I0409 16:52:59.159163 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.823389 | |
| I0409 16:52:59.159175 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.954364 | |
| I0409 16:52:59.159188 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.91312 | |
| I0409 16:52:59.159201 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.724724 (* 1 = 0.724724 loss) | |
| I0409 16:52:59.159215 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.194869 (* 1 = 0.194869 loss) | |
| I0409 16:52:59.159227 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.534 | |
| I0409 16:52:59.159238 12249 solver.cpp:406] Test net output #16: total_confidence = 0.487336 | |
| I0409 16:52:59.535024 12249 solver.cpp:229] Iteration 175000, loss = 2.27472 | |
| I0409 16:52:59.535089 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.571429 | |
| I0409 16:52:59.535106 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.880682 | |
| I0409 16:52:59.535120 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.809524 | |
| I0409 16:52:59.535136 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.45453 (* 0.3 = 0.436359 loss) | |
| I0409 16:52:59.535151 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.401233 (* 0.3 = 0.12037 loss) | |
| I0409 16:52:59.535164 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.714286 | |
| I0409 16:52:59.535176 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.920455 | |
| I0409 16:52:59.535188 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.880952 | |
| I0409 16:52:59.535203 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.15793 (* 0.3 = 0.34738 loss) | |
| I0409 16:52:59.535218 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.318287 (* 0.3 = 0.095486 loss) | |
| I0409 16:52:59.535230 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.880952 | |
| I0409 16:52:59.535243 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 16:52:59.535254 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.952381 | |
| I0409 16:52:59.535269 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.530678 (* 1 = 0.530678 loss) | |
| I0409 16:52:59.535282 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.164988 (* 1 = 0.164988 loss) | |
| I0409 16:52:59.535295 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 16:52:59.535307 12249 solver.cpp:245] Train net output #16: total_confidence = 0.598115 | |
| I0409 16:52:59.535321 12249 sgd_solver.cpp:106] Iteration 175000, lr = 0.0075 | |
| I0409 16:58:32.955554 12249 solver.cpp:229] Iteration 175500, loss = 2.25388 | |
| I0409 16:58:32.955716 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.311111 | |
| I0409 16:58:32.955737 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 16:58:32.955754 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.511111 | |
| I0409 16:58:32.955770 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.4495 (* 0.3 = 0.73485 loss) | |
| I0409 16:58:32.955785 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.669101 (* 0.3 = 0.20073 loss) | |
| I0409 16:58:32.955798 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.533333 | |
| I0409 16:58:32.955811 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 16:58:32.955822 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.844444 | |
| I0409 16:58:32.955837 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.60506 (* 0.3 = 0.481519 loss) | |
| I0409 16:58:32.955850 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.452922 (* 0.3 = 0.135877 loss) | |
| I0409 16:58:32.955862 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.822222 | |
| I0409 16:58:32.955874 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 16:58:32.955886 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.911111 | |
| I0409 16:58:32.955900 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.14634 (* 1 = 1.14634 loss) | |
| I0409 16:58:32.955914 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.311389 (* 1 = 0.311389 loss) | |
| I0409 16:58:32.955926 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 16:58:32.955938 12249 solver.cpp:245] Train net output #16: total_confidence = 0.388467 | |
| I0409 16:58:32.955953 12249 sgd_solver.cpp:106] Iteration 175500, lr = 0.00749286 | |
| I0409 17:04:06.335975 12249 solver.cpp:229] Iteration 176000, loss = 2.26146 | |
| I0409 17:04:06.336316 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.428571 | |
| I0409 17:04:06.336338 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 17:04:06.336350 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.642857 | |
| I0409 17:04:06.336367 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.93617 (* 0.3 = 0.58085 loss) | |
| I0409 17:04:06.336385 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.492267 (* 0.3 = 0.14768 loss) | |
| I0409 17:04:06.336396 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.571429 | |
| I0409 17:04:06.336408 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 17:04:06.336421 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.833333 | |
| I0409 17:04:06.336434 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.28127 (* 0.3 = 0.384382 loss) | |
| I0409 17:04:06.336448 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.336683 (* 0.3 = 0.101005 loss) | |
| I0409 17:04:06.336462 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.904762 | |
| I0409 17:04:06.336473 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 17:04:06.336506 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.952381 | |
| I0409 17:04:06.336524 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.303643 (* 1 = 0.303643 loss) | |
| I0409 17:04:06.336539 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.078732 (* 1 = 0.078732 loss) | |
| I0409 17:04:06.336550 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 17:04:06.336562 12249 solver.cpp:245] Train net output #16: total_confidence = 0.529136 | |
| I0409 17:04:06.336577 12249 sgd_solver.cpp:106] Iteration 176000, lr = 0.00748571 | |
| I0409 17:09:39.714808 12249 solver.cpp:229] Iteration 176500, loss = 2.23284 | |
| I0409 17:09:39.714951 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.533333 | |
| I0409 17:09:39.714978 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 17:09:39.715001 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.822222 | |
| I0409 17:09:39.715029 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.33525 (* 0.3 = 0.400576 loss) | |
| I0409 17:09:39.715055 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.405679 (* 0.3 = 0.121704 loss) | |
| I0409 17:09:39.715078 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.711111 | |
| I0409 17:09:39.715101 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 17:09:39.715124 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.955556 | |
| I0409 17:09:39.715148 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.891324 (* 0.3 = 0.267397 loss) | |
| I0409 17:09:39.715174 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.297048 (* 0.3 = 0.0891145 loss) | |
| I0409 17:09:39.715198 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.933333 | |
| I0409 17:09:39.715217 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 17:09:39.715240 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 17:09:39.715266 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.255608 (* 1 = 0.255608 loss) | |
| I0409 17:09:39.715293 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0738675 (* 1 = 0.0738675 loss) | |
| I0409 17:09:39.715315 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 17:09:39.715337 12249 solver.cpp:245] Train net output #16: total_confidence = 0.462178 | |
| I0409 17:09:39.715361 12249 sgd_solver.cpp:106] Iteration 176500, lr = 0.00747857 | |
| I0409 17:15:13.086069 12249 solver.cpp:229] Iteration 177000, loss = 2.21173 | |
| I0409 17:15:13.086360 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.259259 | |
| I0409 17:15:13.086380 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.761364 | |
| I0409 17:15:13.086393 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.592593 | |
| I0409 17:15:13.086410 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.32307 (* 0.3 = 0.696922 loss) | |
| I0409 17:15:13.086426 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.747929 (* 0.3 = 0.224379 loss) | |
| I0409 17:15:13.086438 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 17:15:13.086452 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 17:15:13.086463 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.740741 | |
| I0409 17:15:13.086477 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.87264 (* 0.3 = 0.561792 loss) | |
| I0409 17:15:13.086491 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.588957 (* 0.3 = 0.176687 loss) | |
| I0409 17:15:13.086504 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.759259 | |
| I0409 17:15:13.086516 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 17:15:13.086529 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.962963 | |
| I0409 17:15:13.086542 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.963783 (* 1 = 0.963783 loss) | |
| I0409 17:15:13.086557 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.29992 (* 1 = 0.29992 loss) | |
| I0409 17:15:13.086570 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 17:15:13.086581 12249 solver.cpp:245] Train net output #16: total_confidence = 0.323042 | |
| I0409 17:15:13.086596 12249 sgd_solver.cpp:106] Iteration 177000, lr = 0.00747143 | |
| I0409 17:20:46.448889 12249 solver.cpp:229] Iteration 177500, loss = 2.21142 | |
| I0409 17:20:46.449010 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.425532 | |
| I0409 17:20:46.449030 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.840909 | |
| I0409 17:20:46.449043 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.765957 | |
| I0409 17:20:46.449060 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.73261 (* 0.3 = 0.519784 loss) | |
| I0409 17:20:46.449075 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.480514 (* 0.3 = 0.144154 loss) | |
| I0409 17:20:46.449087 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.638298 | |
| I0409 17:20:46.449100 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 17:20:46.449111 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.829787 | |
| I0409 17:20:46.449126 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.26264 (* 0.3 = 0.378792 loss) | |
| I0409 17:20:46.449141 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.3611 (* 0.3 = 0.10833 loss) | |
| I0409 17:20:46.449153 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.93617 | |
| I0409 17:20:46.449165 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 17:20:46.449177 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.978723 | |
| I0409 17:20:46.449192 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.352459 (* 1 = 0.352459 loss) | |
| I0409 17:20:46.449206 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0967293 (* 1 = 0.0967293 loss) | |
| I0409 17:20:46.449218 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 17:20:46.449230 12249 solver.cpp:245] Train net output #16: total_confidence = 0.535571 | |
| I0409 17:20:46.449245 12249 sgd_solver.cpp:106] Iteration 177500, lr = 0.00746429 | |
| I0409 17:26:19.828718 12249 solver.cpp:229] Iteration 178000, loss = 2.21916 | |
| I0409 17:26:19.829089 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.410256 | |
| I0409 17:26:19.829112 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 17:26:19.829125 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.641026 | |
| I0409 17:26:19.829143 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.0768 (* 0.3 = 0.62304 loss) | |
| I0409 17:26:19.829157 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.560124 (* 0.3 = 0.168037 loss) | |
| I0409 17:26:19.829170 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.692308 | |
| I0409 17:26:19.829182 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.914773 | |
| I0409 17:26:19.829195 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.871795 | |
| I0409 17:26:19.829210 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.10719 (* 0.3 = 0.332157 loss) | |
| I0409 17:26:19.829223 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.318543 (* 0.3 = 0.095563 loss) | |
| I0409 17:26:19.829236 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.897436 | |
| I0409 17:26:19.829248 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 17:26:19.829262 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.948718 | |
| I0409 17:26:19.829275 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.406235 (* 1 = 0.406235 loss) | |
| I0409 17:26:19.829289 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.114291 (* 1 = 0.114291 loss) | |
| I0409 17:26:19.829303 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 17:26:19.829314 12249 solver.cpp:245] Train net output #16: total_confidence = 0.388592 | |
| I0409 17:26:19.829329 12249 sgd_solver.cpp:106] Iteration 178000, lr = 0.00745714 | |
| I0409 17:31:53.200016 12249 solver.cpp:229] Iteration 178500, loss = 2.27778 | |
| I0409 17:31:53.200276 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.46 | |
| I0409 17:31:53.200307 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 17:31:53.200332 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.7 | |
| I0409 17:31:53.200361 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.04358 (* 0.3 = 0.613074 loss) | |
| I0409 17:31:53.200381 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.643741 (* 0.3 = 0.193122 loss) | |
| I0409 17:31:53.200397 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.72 | |
| I0409 17:31:53.200409 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.920455 | |
| I0409 17:31:53.200422 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.88 | |
| I0409 17:31:53.200435 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.20405 (* 0.3 = 0.361215 loss) | |
| I0409 17:31:53.200450 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.359464 (* 0.3 = 0.107839 loss) | |
| I0409 17:31:53.200462 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.86 | |
| I0409 17:31:53.200474 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 17:31:53.200501 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.96 | |
| I0409 17:31:53.200518 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.443728 (* 1 = 0.443728 loss) | |
| I0409 17:31:53.200532 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.148677 (* 1 = 0.148677 loss) | |
| I0409 17:31:53.200544 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 17:31:53.200556 12249 solver.cpp:245] Train net output #16: total_confidence = 0.393436 | |
| I0409 17:31:53.200572 12249 sgd_solver.cpp:106] Iteration 178500, lr = 0.00745 | |
| I0409 17:37:26.567981 12249 solver.cpp:229] Iteration 179000, loss = 2.2338 | |
| I0409 17:37:26.568140 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.355556 | |
| I0409 17:37:26.568169 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 17:37:26.568192 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.622222 | |
| I0409 17:37:26.568222 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.23917 (* 0.3 = 0.671752 loss) | |
| I0409 17:37:26.568249 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.633498 (* 0.3 = 0.19005 loss) | |
| I0409 17:37:26.568271 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.511111 | |
| I0409 17:37:26.568292 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 17:37:26.568312 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.822222 | |
| I0409 17:37:26.568341 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.6159 (* 0.3 = 0.484769 loss) | |
| I0409 17:37:26.568367 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.459861 (* 0.3 = 0.137958 loss) | |
| I0409 17:37:26.568388 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.688889 | |
| I0409 17:37:26.568409 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.920455 | |
| I0409 17:37:26.568430 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.866667 | |
| I0409 17:37:26.568455 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.0485 (* 1 = 1.0485 loss) | |
| I0409 17:37:26.568498 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.280955 (* 1 = 0.280955 loss) | |
| I0409 17:37:26.568524 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 17:37:26.568547 12249 solver.cpp:245] Train net output #16: total_confidence = 0.296686 | |
| I0409 17:37:26.568570 12249 sgd_solver.cpp:106] Iteration 179000, lr = 0.00744286 | |
| I0409 17:43:00.297263 12249 solver.cpp:229] Iteration 179500, loss = 2.2287 | |
| I0409 17:43:00.297503 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.372093 | |
| I0409 17:43:00.297524 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 17:43:00.297538 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.55814 | |
| I0409 17:43:00.297554 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.46683 (* 0.3 = 0.74005 loss) | |
| I0409 17:43:00.297569 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.691899 (* 0.3 = 0.20757 loss) | |
| I0409 17:43:00.297581 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.488372 | |
| I0409 17:43:00.297593 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 17:43:00.297605 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72093 | |
| I0409 17:43:00.297619 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.93874 (* 0.3 = 0.581622 loss) | |
| I0409 17:43:00.297633 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.543178 (* 0.3 = 0.162954 loss) | |
| I0409 17:43:00.297646 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.697674 | |
| I0409 17:43:00.297658 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 17:43:00.297670 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.767442 | |
| I0409 17:43:00.297684 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.52129 (* 1 = 1.52129 loss) | |
| I0409 17:43:00.297698 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.420224 (* 1 = 0.420224 loss) | |
| I0409 17:43:00.297711 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 17:43:00.297724 12249 solver.cpp:245] Train net output #16: total_confidence = 0.367608 | |
| I0409 17:43:00.297739 12249 sgd_solver.cpp:106] Iteration 179500, lr = 0.00743571 | |
| I0409 17:48:33.258967 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_180000.caffemodel | |
| I0409 17:48:33.745582 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_180000.solverstate | |
| I0409 17:48:33.993926 12249 solver.cpp:338] Iteration 180000, Testing net (#0) | |
| I0409 17:49:15.018455 12249 solver.cpp:393] Test loss: 1.93503 | |
| I0409 17:49:15.018571 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.503377 | |
| I0409 17:49:15.018591 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.873094 | |
| I0409 17:49:15.018604 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.792699 | |
| I0409 17:49:15.018621 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.70533 (* 0.3 = 0.5116 loss) | |
| I0409 17:49:15.018636 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.441274 (* 0.3 = 0.132382 loss) | |
| I0409 17:49:15.018648 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.704162 | |
| I0409 17:49:15.018661 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.923638 | |
| I0409 17:49:15.018672 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.893033 | |
| I0409 17:49:15.018687 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.0691 (* 0.3 = 0.320731 loss) | |
| I0409 17:49:15.018700 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.278649 (* 0.3 = 0.0835948 loss) | |
| I0409 17:49:15.018712 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.834521 | |
| I0409 17:49:15.018724 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.960137 | |
| I0409 17:49:15.018736 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.922212 | |
| I0409 17:49:15.018759 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.707285 (* 1 = 0.707285 loss) | |
| I0409 17:49:15.018784 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.179436 (* 1 = 0.179436 loss) | |
| I0409 17:49:15.018797 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.586 | |
| I0409 17:49:15.018808 12249 solver.cpp:406] Test net output #16: total_confidence = 0.507133 | |
| I0409 17:49:15.391975 12249 solver.cpp:229] Iteration 180000, loss = 2.23578 | |
| I0409 17:49:15.392040 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.477273 | |
| I0409 17:49:15.392057 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 17:49:15.392071 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.727273 | |
| I0409 17:49:15.392087 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.75292 (* 0.3 = 0.525877 loss) | |
| I0409 17:49:15.392102 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.485661 (* 0.3 = 0.145698 loss) | |
| I0409 17:49:15.392115 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.704545 | |
| I0409 17:49:15.392127 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 17:49:15.392140 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.909091 | |
| I0409 17:49:15.392154 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.841706 (* 0.3 = 0.252512 loss) | |
| I0409 17:49:15.392169 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.264084 (* 0.3 = 0.0792251 loss) | |
| I0409 17:49:15.392181 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.886364 | |
| I0409 17:49:15.392194 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 17:49:15.392205 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.954545 | |
| I0409 17:49:15.392223 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.348649 (* 1 = 0.348649 loss) | |
| I0409 17:49:15.392238 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0949748 (* 1 = 0.0949748 loss) | |
| I0409 17:49:15.392251 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 17:49:15.392263 12249 solver.cpp:245] Train net output #16: total_confidence = 0.511888 | |
| I0409 17:49:15.392278 12249 sgd_solver.cpp:106] Iteration 180000, lr = 0.00742857 | |
| I0409 17:54:48.698468 12249 solver.cpp:229] Iteration 180500, loss = 2.22692 | |
| I0409 17:54:48.698797 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.348837 | |
| I0409 17:54:48.698818 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 17:54:48.698832 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.767442 | |
| I0409 17:54:48.698848 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.17782 (* 0.3 = 0.653345 loss) | |
| I0409 17:54:48.698863 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.642007 (* 0.3 = 0.192602 loss) | |
| I0409 17:54:48.698876 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.418605 | |
| I0409 17:54:48.698889 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 17:54:48.698900 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.72093 | |
| I0409 17:54:48.698915 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.7545 (* 0.3 = 0.526351 loss) | |
| I0409 17:54:48.698928 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.558752 (* 0.3 = 0.167626 loss) | |
| I0409 17:54:48.698940 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.790698 | |
| I0409 17:54:48.698952 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 17:54:48.698964 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.953488 | |
| I0409 17:54:48.698979 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.767564 (* 1 = 0.767564 loss) | |
| I0409 17:54:48.698993 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.257435 (* 1 = 0.257435 loss) | |
| I0409 17:54:48.699005 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 17:54:48.699018 12249 solver.cpp:245] Train net output #16: total_confidence = 0.328296 | |
| I0409 17:54:48.699031 12249 sgd_solver.cpp:106] Iteration 180500, lr = 0.00742143 | |
| I0409 18:00:22.081393 12249 solver.cpp:229] Iteration 181000, loss = 2.21416 | |
| I0409 18:00:22.081511 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.489362 | |
| I0409 18:00:22.081529 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 18:00:22.081543 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.765957 | |
| I0409 18:00:22.081560 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.54094 (* 0.3 = 0.462282 loss) | |
| I0409 18:00:22.081575 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.447086 (* 0.3 = 0.134126 loss) | |
| I0409 18:00:22.081588 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.595745 | |
| I0409 18:00:22.081600 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 18:00:22.081614 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.957447 | |
| I0409 18:00:22.081627 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.10523 (* 0.3 = 0.33157 loss) | |
| I0409 18:00:22.081642 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.33862 (* 0.3 = 0.101586 loss) | |
| I0409 18:00:22.081655 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.93617 | |
| I0409 18:00:22.081667 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 18:00:22.081679 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 18:00:22.081693 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.230118 (* 1 = 0.230118 loss) | |
| I0409 18:00:22.081708 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0700585 (* 1 = 0.0700585 loss) | |
| I0409 18:00:22.081720 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 18:00:22.081733 12249 solver.cpp:245] Train net output #16: total_confidence = 0.485238 | |
| I0409 18:00:22.081748 12249 sgd_solver.cpp:106] Iteration 181000, lr = 0.00741429 | |
| I0409 18:05:55.429587 12249 solver.cpp:229] Iteration 181500, loss = 2.25535 | |
| I0409 18:05:55.429956 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.227273 | |
| I0409 18:05:55.429980 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 18:05:55.429992 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.568182 | |
| I0409 18:05:55.430009 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.34851 (* 0.3 = 0.704553 loss) | |
| I0409 18:05:55.430025 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.65255 (* 0.3 = 0.195765 loss) | |
| I0409 18:05:55.430038 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.636364 | |
| I0409 18:05:55.430050 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 18:05:55.430063 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.818182 | |
| I0409 18:05:55.430076 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.39855 (* 0.3 = 0.419564 loss) | |
| I0409 18:05:55.430091 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.394172 (* 0.3 = 0.118252 loss) | |
| I0409 18:05:55.430104 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.931818 | |
| I0409 18:05:55.430115 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.982955 | |
| I0409 18:05:55.430127 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.954545 | |
| I0409 18:05:55.430142 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.485901 (* 1 = 0.485901 loss) | |
| I0409 18:05:55.430156 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.135564 (* 1 = 0.135564 loss) | |
| I0409 18:05:55.430168 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 18:05:55.430181 12249 solver.cpp:245] Train net output #16: total_confidence = 0.408593 | |
| I0409 18:05:55.430197 12249 sgd_solver.cpp:106] Iteration 181500, lr = 0.00740714 | |
| I0409 18:11:28.792356 12249 solver.cpp:229] Iteration 182000, loss = 2.23339 | |
| I0409 18:11:28.792469 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.387755 | |
| I0409 18:11:28.792505 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 18:11:28.792520 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.857143 | |
| I0409 18:11:28.792537 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.60051 (* 0.3 = 0.480153 loss) | |
| I0409 18:11:28.792552 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.460655 (* 0.3 = 0.138197 loss) | |
| I0409 18:11:28.792565 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.632653 | |
| I0409 18:11:28.792578 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 18:11:28.792590 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.918367 | |
| I0409 18:11:28.792604 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.09283 (* 0.3 = 0.32785 loss) | |
| I0409 18:11:28.792618 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.324208 (* 0.3 = 0.0972623 loss) | |
| I0409 18:11:28.792631 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.938776 | |
| I0409 18:11:28.792644 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 18:11:28.792656 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 18:11:28.792670 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.168774 (* 1 = 0.168774 loss) | |
| I0409 18:11:28.792686 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.057426 (* 1 = 0.057426 loss) | |
| I0409 18:11:28.792700 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 18:11:28.792711 12249 solver.cpp:245] Train net output #16: total_confidence = 0.521803 | |
| I0409 18:11:28.792726 12249 sgd_solver.cpp:106] Iteration 182000, lr = 0.0074 | |
| I0409 18:17:02.173941 12249 solver.cpp:229] Iteration 182500, loss = 2.22953 | |
| I0409 18:17:02.174300 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4375 | |
| I0409 18:17:02.174324 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 18:17:02.174337 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.625 | |
| I0409 18:17:02.174353 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.31978 (* 0.3 = 0.695935 loss) | |
| I0409 18:17:02.174370 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.693886 (* 0.3 = 0.208166 loss) | |
| I0409 18:17:02.174382 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.583333 | |
| I0409 18:17:02.174394 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 18:17:02.174407 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.791667 | |
| I0409 18:17:02.174422 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.51408 (* 0.3 = 0.454225 loss) | |
| I0409 18:17:02.174437 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.46378 (* 0.3 = 0.139134 loss) | |
| I0409 18:17:02.174449 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.875 | |
| I0409 18:17:02.174461 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 18:17:02.174474 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.895833 | |
| I0409 18:17:02.174489 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.687994 (* 1 = 0.687994 loss) | |
| I0409 18:17:02.174504 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.206273 (* 1 = 0.206273 loss) | |
| I0409 18:17:02.174516 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 18:17:02.174528 12249 solver.cpp:245] Train net output #16: total_confidence = 0.456017 | |
| I0409 18:17:02.174543 12249 sgd_solver.cpp:106] Iteration 182500, lr = 0.00739286 | |
| I0409 18:22:35.545428 12249 solver.cpp:229] Iteration 183000, loss = 2.20773 | |
| I0409 18:22:35.545737 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.306122 | |
| I0409 18:22:35.545758 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 18:22:35.545771 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.612245 | |
| I0409 18:22:35.545789 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.36963 (* 0.3 = 0.710888 loss) | |
| I0409 18:22:35.545804 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.685183 (* 0.3 = 0.205555 loss) | |
| I0409 18:22:35.545820 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.530612 | |
| I0409 18:22:35.545835 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 18:22:35.545846 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.755102 | |
| I0409 18:22:35.545861 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.64366 (* 0.3 = 0.493097 loss) | |
| I0409 18:22:35.545874 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.496828 (* 0.3 = 0.149048 loss) | |
| I0409 18:22:35.545887 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.795918 | |
| I0409 18:22:35.545899 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 18:22:35.545912 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.877551 | |
| I0409 18:22:35.545925 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.872403 (* 1 = 0.872403 loss) | |
| I0409 18:22:35.545940 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.24799 (* 1 = 0.24799 loss) | |
| I0409 18:22:35.545953 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 18:22:35.545965 12249 solver.cpp:245] Train net output #16: total_confidence = 0.370731 | |
| I0409 18:22:35.545980 12249 sgd_solver.cpp:106] Iteration 183000, lr = 0.00738571 | |
| I0409 18:28:08.930145 12249 solver.cpp:229] Iteration 183500, loss = 2.23015 | |
| I0409 18:28:08.930295 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.521739 | |
| I0409 18:28:08.930316 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 18:28:08.930330 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.73913 | |
| I0409 18:28:08.930346 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.65106 (* 0.3 = 0.495318 loss) | |
| I0409 18:28:08.930361 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.490848 (* 0.3 = 0.147254 loss) | |
| I0409 18:28:08.930374 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.717391 | |
| I0409 18:28:08.930387 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 18:28:08.930398 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.913043 | |
| I0409 18:28:08.930413 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.83974 (* 0.3 = 0.251922 loss) | |
| I0409 18:28:08.930428 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.263932 (* 0.3 = 0.0791795 loss) | |
| I0409 18:28:08.930439 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.826087 | |
| I0409 18:28:08.930451 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 18:28:08.930464 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.869565 | |
| I0409 18:28:08.930479 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.951998 (* 1 = 0.951998 loss) | |
| I0409 18:28:08.930492 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.292312 (* 1 = 0.292312 loss) | |
| I0409 18:28:08.930505 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 18:28:08.930516 12249 solver.cpp:245] Train net output #16: total_confidence = 0.60049 | |
| I0409 18:28:08.930531 12249 sgd_solver.cpp:106] Iteration 183500, lr = 0.00737857 | |
| I0409 18:33:42.283906 12249 solver.cpp:229] Iteration 184000, loss = 2.21676 | |
| I0409 18:33:42.284190 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.513514 | |
| I0409 18:33:42.284219 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.863636 | |
| I0409 18:33:42.284242 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.648649 | |
| I0409 18:33:42.284271 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.75044 (* 0.3 = 0.525133 loss) | |
| I0409 18:33:42.284297 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.495296 (* 0.3 = 0.148589 loss) | |
| I0409 18:33:42.284319 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.756757 | |
| I0409 18:33:42.284343 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 18:33:42.284365 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.837838 | |
| I0409 18:33:42.284394 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.12061 (* 0.3 = 0.336183 loss) | |
| I0409 18:33:42.284420 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.439302 (* 0.3 = 0.131791 loss) | |
| I0409 18:33:42.284441 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.783784 | |
| I0409 18:33:42.284462 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 18:33:42.284504 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.864865 | |
| I0409 18:33:42.284535 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.693767 (* 1 = 0.693767 loss) | |
| I0409 18:33:42.284562 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.165193 (* 1 = 0.165193 loss) | |
| I0409 18:33:42.284584 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 18:33:42.284605 12249 solver.cpp:245] Train net output #16: total_confidence = 0.409432 | |
| I0409 18:33:42.284631 12249 sgd_solver.cpp:106] Iteration 184000, lr = 0.00737143 | |
| I0409 18:39:15.661748 12249 solver.cpp:229] Iteration 184500, loss = 2.20419 | |
| I0409 18:39:15.661898 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.236842 | |
| I0409 18:39:15.661918 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 18:39:15.661931 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.473684 | |
| I0409 18:39:15.661948 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 3.12114 (* 0.3 = 0.936342 loss) | |
| I0409 18:39:15.661964 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.790871 (* 0.3 = 0.237261 loss) | |
| I0409 18:39:15.661977 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.5 | |
| I0409 18:39:15.661989 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 18:39:15.662001 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.605263 | |
| I0409 18:39:15.662015 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.22295 (* 0.3 = 0.666884 loss) | |
| I0409 18:39:15.662029 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.552968 (* 0.3 = 0.165891 loss) | |
| I0409 18:39:15.662041 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.526316 | |
| I0409 18:39:15.662053 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.886364 | |
| I0409 18:39:15.662066 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.657895 | |
| I0409 18:39:15.662081 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 2.05775 (* 1 = 2.05775 loss) | |
| I0409 18:39:15.662096 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.487226 (* 1 = 0.487226 loss) | |
| I0409 18:39:15.662107 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 18:39:15.662119 12249 solver.cpp:245] Train net output #16: total_confidence = 0.337309 | |
| I0409 18:39:15.662134 12249 sgd_solver.cpp:106] Iteration 184500, lr = 0.00736429 | |
| I0409 18:44:48.634768 12249 solver.cpp:338] Iteration 185000, Testing net (#0) | |
| I0409 18:45:30.022099 12249 solver.cpp:393] Test loss: 1.94773 | |
| I0409 18:45:30.022187 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.536832 | |
| I0409 18:45:30.022205 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.875367 | |
| I0409 18:45:30.022218 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.80797 | |
| I0409 18:45:30.022235 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.59603 (* 0.3 = 0.478808 loss) | |
| I0409 18:45:30.022250 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.438559 (* 0.3 = 0.131568 loss) | |
| I0409 18:45:30.022263 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.714587 | |
| I0409 18:45:30.022274 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.920683 | |
| I0409 18:45:30.022287 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.899041 | |
| I0409 18:45:30.022300 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.05572 (* 0.3 = 0.316715 loss) | |
| I0409 18:45:30.022315 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.296048 (* 0.3 = 0.0888145 loss) | |
| I0409 18:45:30.022327 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.833849 | |
| I0409 18:45:30.022339 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.958091 | |
| I0409 18:45:30.022351 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.917112 | |
| I0409 18:45:30.022364 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.739657 (* 1 = 0.739657 loss) | |
| I0409 18:45:30.022377 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.19217 (* 1 = 0.19217 loss) | |
| I0409 18:45:30.022389 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.578 | |
| I0409 18:45:30.022402 12249 solver.cpp:406] Test net output #16: total_confidence = 0.534147 | |
| I0409 18:45:30.400782 12249 solver.cpp:229] Iteration 185000, loss = 2.1659 | |
| I0409 18:45:30.400838 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.452381 | |
| I0409 18:45:30.400856 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 18:45:30.400869 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.785714 | |
| I0409 18:45:30.400887 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.547 (* 0.3 = 0.464099 loss) | |
| I0409 18:45:30.400902 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.472812 (* 0.3 = 0.141844 loss) | |
| I0409 18:45:30.400914 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.642857 | |
| I0409 18:45:30.400928 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.892045 | |
| I0409 18:45:30.400938 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.880952 | |
| I0409 18:45:30.400952 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.11947 (* 0.3 = 0.33584 loss) | |
| I0409 18:45:30.400967 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.348983 (* 0.3 = 0.104695 loss) | |
| I0409 18:45:30.400979 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.928571 | |
| I0409 18:45:30.400992 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 18:45:30.401005 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.97619 | |
| I0409 18:45:30.401018 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.256266 (* 1 = 0.256266 loss) | |
| I0409 18:45:30.401032 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.160456 (* 1 = 0.160456 loss) | |
| I0409 18:45:30.401044 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 18:45:30.401057 12249 solver.cpp:245] Train net output #16: total_confidence = 0.613299 | |
| I0409 18:45:30.401072 12249 sgd_solver.cpp:106] Iteration 185000, lr = 0.00735714 | |
| I0409 18:51:03.762104 12249 solver.cpp:229] Iteration 185500, loss = 2.19719 | |
| I0409 18:51:03.762284 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.425 | |
| I0409 18:51:03.762305 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 18:51:03.762317 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.8 | |
| I0409 18:51:03.762334 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.47774 (* 0.3 = 0.443323 loss) | |
| I0409 18:51:03.762351 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.374554 (* 0.3 = 0.112366 loss) | |
| I0409 18:51:03.762363 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.75 | |
| I0409 18:51:03.762375 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 18:51:03.762387 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 1 | |
| I0409 18:51:03.762401 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.736459 (* 0.3 = 0.220938 loss) | |
| I0409 18:51:03.762416 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.251146 (* 0.3 = 0.0753438 loss) | |
| I0409 18:51:03.762428 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.975 | |
| I0409 18:51:03.762440 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.994318 | |
| I0409 18:51:03.762452 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 18:51:03.762466 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.0676004 (* 1 = 0.0676004 loss) | |
| I0409 18:51:03.762481 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0255706 (* 1 = 0.0255706 loss) | |
| I0409 18:51:03.762493 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 18:51:03.762506 12249 solver.cpp:245] Train net output #16: total_confidence = 0.594963 | |
| I0409 18:51:03.762521 12249 sgd_solver.cpp:106] Iteration 185500, lr = 0.00735 | |
| I0409 18:56:37.434100 12249 solver.cpp:229] Iteration 186000, loss = 2.19424 | |
| I0409 18:56:37.434422 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.321429 | |
| I0409 18:56:37.434443 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.784091 | |
| I0409 18:56:37.434456 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.642857 | |
| I0409 18:56:37.434473 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.09281 (* 0.3 = 0.627844 loss) | |
| I0409 18:56:37.434489 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.693857 (* 0.3 = 0.208157 loss) | |
| I0409 18:56:37.434500 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.464286 | |
| I0409 18:56:37.434514 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 18:56:37.434525 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.767857 | |
| I0409 18:56:37.434540 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.65466 (* 0.3 = 0.496399 loss) | |
| I0409 18:56:37.434553 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.540642 (* 0.3 = 0.162193 loss) | |
| I0409 18:56:37.434566 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.785714 | |
| I0409 18:56:37.434577 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 18:56:37.434589 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.928571 | |
| I0409 18:56:37.434603 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.777632 (* 1 = 0.777632 loss) | |
| I0409 18:56:37.434617 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.264157 (* 1 = 0.264157 loss) | |
| I0409 18:56:37.434629 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 18:56:37.434643 12249 solver.cpp:245] Train net output #16: total_confidence = 0.306564 | |
| I0409 18:56:37.434656 12249 sgd_solver.cpp:106] Iteration 186000, lr = 0.00734286 | |
| I0409 19:02:10.819536 12249 solver.cpp:229] Iteration 186500, loss = 2.18916 | |
| I0409 19:02:10.819803 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.355556 | |
| I0409 19:02:10.819823 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 19:02:10.819836 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.644444 | |
| I0409 19:02:10.819852 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.06702 (* 0.3 = 0.620107 loss) | |
| I0409 19:02:10.819869 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.584261 (* 0.3 = 0.175278 loss) | |
| I0409 19:02:10.819881 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.666667 | |
| I0409 19:02:10.819893 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 19:02:10.819905 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.866667 | |
| I0409 19:02:10.819921 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.1635 (* 0.3 = 0.34905 loss) | |
| I0409 19:02:10.819934 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.336425 (* 0.3 = 0.100927 loss) | |
| I0409 19:02:10.819947 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.844444 | |
| I0409 19:02:10.819959 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 19:02:10.819973 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.955556 | |
| I0409 19:02:10.819988 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.394688 (* 1 = 0.394688 loss) | |
| I0409 19:02:10.820001 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.104594 (* 1 = 0.104594 loss) | |
| I0409 19:02:10.820014 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 19:02:10.820026 12249 solver.cpp:245] Train net output #16: total_confidence = 0.416137 | |
| I0409 19:02:10.820040 12249 sgd_solver.cpp:106] Iteration 186500, lr = 0.00733571 | |
| I0409 19:07:44.171347 12249 solver.cpp:229] Iteration 187000, loss = 2.16506 | |
| I0409 19:07:44.171562 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.288462 | |
| I0409 19:07:44.171583 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 19:07:44.171597 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.423077 | |
| I0409 19:07:44.171613 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.72656 (* 0.3 = 0.817968 loss) | |
| I0409 19:07:44.171629 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.874631 (* 0.3 = 0.262389 loss) | |
| I0409 19:07:44.171641 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.365385 | |
| I0409 19:07:44.171654 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.795455 | |
| I0409 19:07:44.171666 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.692308 | |
| I0409 19:07:44.171680 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.97084 (* 0.3 = 0.591253 loss) | |
| I0409 19:07:44.171694 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.647364 (* 0.3 = 0.194209 loss) | |
| I0409 19:07:44.171706 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.653846 | |
| I0409 19:07:44.171720 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.880682 | |
| I0409 19:07:44.171731 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.807692 | |
| I0409 19:07:44.171747 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.28998 (* 1 = 1.28998 loss) | |
| I0409 19:07:44.171763 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.418701 (* 1 = 0.418701 loss) | |
| I0409 19:07:44.171775 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 19:07:44.171787 12249 solver.cpp:245] Train net output #16: total_confidence = 0.262176 | |
| I0409 19:07:44.171802 12249 sgd_solver.cpp:106] Iteration 187000, lr = 0.00732857 | |
| I0409 19:13:17.545014 12249 solver.cpp:229] Iteration 187500, loss = 2.1759 | |
| I0409 19:13:17.545289 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.413043 | |
| I0409 19:13:17.545307 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 19:13:17.545320 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.586957 | |
| I0409 19:13:17.545337 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.00906 (* 0.3 = 0.602719 loss) | |
| I0409 19:13:17.545352 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.610436 (* 0.3 = 0.183131 loss) | |
| I0409 19:13:17.545366 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.565217 | |
| I0409 19:13:17.545382 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.869318 | |
| I0409 19:13:17.545394 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.804348 | |
| I0409 19:13:17.545408 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.33074 (* 0.3 = 0.399223 loss) | |
| I0409 19:13:17.545423 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.406178 (* 0.3 = 0.121853 loss) | |
| I0409 19:13:17.545435 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.73913 | |
| I0409 19:13:17.545447 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.909091 | |
| I0409 19:13:17.545459 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 19:13:17.545476 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.742441 (* 1 = 0.742441 loss) | |
| I0409 19:13:17.545491 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.262528 (* 1 = 0.262528 loss) | |
| I0409 19:13:17.545505 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 19:13:17.545517 12249 solver.cpp:245] Train net output #16: total_confidence = 0.394946 | |
| I0409 19:13:17.545532 12249 sgd_solver.cpp:106] Iteration 187500, lr = 0.00732143 | |
| I0409 19:18:51.251920 12249 solver.cpp:229] Iteration 188000, loss = 2.20919 | |
| I0409 19:18:51.252071 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.42 | |
| I0409 19:18:51.252092 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 19:18:51.252106 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.74 | |
| I0409 19:18:51.252122 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.80421 (* 0.3 = 0.541264 loss) | |
| I0409 19:18:51.252138 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.543417 (* 0.3 = 0.163025 loss) | |
| I0409 19:18:51.252151 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.64 | |
| I0409 19:18:51.252163 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 19:18:51.252176 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.92 | |
| I0409 19:18:51.252190 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.06903 (* 0.3 = 0.320708 loss) | |
| I0409 19:18:51.252204 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.363971 (* 0.3 = 0.109191 loss) | |
| I0409 19:18:51.252218 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.9 | |
| I0409 19:18:51.252229 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 19:18:51.252241 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 19:18:51.252255 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.25775 (* 1 = 0.25775 loss) | |
| I0409 19:18:51.252270 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0856372 (* 1 = 0.0856372 loss) | |
| I0409 19:18:51.252284 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 19:18:51.252295 12249 solver.cpp:245] Train net output #16: total_confidence = 0.471162 | |
| I0409 19:18:51.252310 12249 sgd_solver.cpp:106] Iteration 188000, lr = 0.00731429 | |
| I0409 19:24:24.979512 12249 solver.cpp:229] Iteration 188500, loss = 2.21316 | |
| I0409 19:24:24.979836 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 19:24:24.979857 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 19:24:24.979871 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.652174 | |
| I0409 19:24:24.979887 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.96444 (* 0.3 = 0.589333 loss) | |
| I0409 19:24:24.979902 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.591058 (* 0.3 = 0.177318 loss) | |
| I0409 19:24:24.979915 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.673913 | |
| I0409 19:24:24.979928 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.886364 | |
| I0409 19:24:24.979939 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.869565 | |
| I0409 19:24:24.979954 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.12129 (* 0.3 = 0.336388 loss) | |
| I0409 19:24:24.979969 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.371927 (* 0.3 = 0.111578 loss) | |
| I0409 19:24:24.979981 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.913043 | |
| I0409 19:24:24.979993 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 19:24:24.980005 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.934783 | |
| I0409 19:24:24.980020 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.589692 (* 1 = 0.589692 loss) | |
| I0409 19:24:24.980034 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.192055 (* 1 = 0.192055 loss) | |
| I0409 19:24:24.980046 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 19:24:24.980058 12249 solver.cpp:245] Train net output #16: total_confidence = 0.55342 | |
| I0409 19:24:24.980072 12249 sgd_solver.cpp:106] Iteration 188500, lr = 0.00730714 | |
| I0409 19:29:58.336009 12249 solver.cpp:229] Iteration 189000, loss = 2.23904 | |
| I0409 19:29:58.336196 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.45283 | |
| I0409 19:29:58.336228 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 19:29:58.336252 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.641509 | |
| I0409 19:29:58.336272 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.09049 (* 0.3 = 0.627148 loss) | |
| I0409 19:29:58.336287 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.693224 (* 0.3 = 0.207967 loss) | |
| I0409 19:29:58.336300 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.490566 | |
| I0409 19:29:58.336313 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.835227 | |
| I0409 19:29:58.336324 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.716981 | |
| I0409 19:29:58.336338 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.62765 (* 0.3 = 0.488296 loss) | |
| I0409 19:29:58.336354 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.529339 (* 0.3 = 0.158802 loss) | |
| I0409 19:29:58.336365 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.811321 | |
| I0409 19:29:58.336377 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.931818 | |
| I0409 19:29:58.336390 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.886792 | |
| I0409 19:29:58.336403 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.67135 (* 1 = 0.67135 loss) | |
| I0409 19:29:58.336417 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.224996 (* 1 = 0.224996 loss) | |
| I0409 19:29:58.336431 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 19:29:58.336443 12249 solver.cpp:245] Train net output #16: total_confidence = 0.333743 | |
| I0409 19:29:58.336457 12249 sgd_solver.cpp:106] Iteration 189000, lr = 0.0073 | |
| I0409 19:35:31.714242 12249 solver.cpp:229] Iteration 189500, loss = 2.17631 | |
| I0409 19:35:31.714542 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.325581 | |
| I0409 19:35:31.714563 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.806818 | |
| I0409 19:35:31.714576 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.511628 | |
| I0409 19:35:31.714594 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.85234 (* 0.3 = 0.855701 loss) | |
| I0409 19:35:31.714609 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.800171 (* 0.3 = 0.240051 loss) | |
| I0409 19:35:31.714622 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.534884 | |
| I0409 19:35:31.714634 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 19:35:31.714646 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.697674 | |
| I0409 19:35:31.714660 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.97383 (* 0.3 = 0.592148 loss) | |
| I0409 19:35:31.714675 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.566286 (* 0.3 = 0.169886 loss) | |
| I0409 19:35:31.714687 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.674419 | |
| I0409 19:35:31.714699 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.892045 | |
| I0409 19:35:31.714711 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.744186 | |
| I0409 19:35:31.714726 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.2362 (* 1 = 1.2362 loss) | |
| I0409 19:35:31.714745 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.372045 (* 1 = 0.372045 loss) | |
| I0409 19:35:31.714756 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 19:35:31.714768 12249 solver.cpp:245] Train net output #16: total_confidence = 0.305544 | |
| I0409 19:35:31.714784 12249 sgd_solver.cpp:106] Iteration 189500, lr = 0.00729286 | |
| I0409 19:41:04.675607 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_190000.caffemodel | |
| I0409 19:41:05.137645 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_190000.solverstate | |
| I0409 19:41:05.380820 12249 solver.cpp:338] Iteration 190000, Testing net (#0) | |
| I0409 19:41:46.462391 12249 solver.cpp:393] Test loss: 1.89503 | |
| I0409 19:41:46.462486 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.516005 | |
| I0409 19:41:46.462503 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.874049 | |
| I0409 19:41:46.462517 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.787693 | |
| I0409 19:41:46.462532 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.70646 (* 0.3 = 0.511938 loss) | |
| I0409 19:41:46.462548 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.450503 (* 0.3 = 0.135151 loss) | |
| I0409 19:41:46.462560 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.735479 | |
| I0409 19:41:46.462573 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.924774 | |
| I0409 19:41:46.462584 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.898292 | |
| I0409 19:41:46.462597 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 0.989079 (* 0.3 = 0.296724 loss) | |
| I0409 19:41:46.462611 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.276665 (* 0.3 = 0.0829994 loss) | |
| I0409 19:41:46.462623 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.835007 | |
| I0409 19:41:46.462635 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.960228 | |
| I0409 19:41:46.462646 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.917118 | |
| I0409 19:41:46.462661 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.692851 (* 1 = 0.692851 loss) | |
| I0409 19:41:46.462674 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.175363 (* 1 = 0.175363 loss) | |
| I0409 19:41:46.462687 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.604 | |
| I0409 19:41:46.462698 12249 solver.cpp:406] Test net output #16: total_confidence = 0.510485 | |
| I0409 19:41:46.835480 12249 solver.cpp:229] Iteration 190000, loss = 2.1684 | |
| I0409 19:41:46.835538 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.294118 | |
| I0409 19:41:46.835556 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.778409 | |
| I0409 19:41:46.835568 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.627451 | |
| I0409 19:41:46.835585 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.3458 (* 0.3 = 0.703741 loss) | |
| I0409 19:41:46.835600 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.730468 (* 0.3 = 0.219141 loss) | |
| I0409 19:41:46.835613 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.411765 | |
| I0409 19:41:46.835626 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.818182 | |
| I0409 19:41:46.835638 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.784314 | |
| I0409 19:41:46.835652 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.85014 (* 0.3 = 0.555041 loss) | |
| I0409 19:41:46.835666 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.557108 (* 0.3 = 0.167132 loss) | |
| I0409 19:41:46.835680 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.705882 | |
| I0409 19:41:46.835691 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 19:41:46.835703 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.862745 | |
| I0409 19:41:46.835717 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.21715 (* 1 = 1.21715 loss) | |
| I0409 19:41:46.835732 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.411558 (* 1 = 0.411558 loss) | |
| I0409 19:41:46.835744 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 19:41:46.835757 12249 solver.cpp:245] Train net output #16: total_confidence = 0.288221 | |
| I0409 19:41:46.835772 12249 sgd_solver.cpp:106] Iteration 190000, lr = 0.00728571 | |
| I0409 19:47:20.113459 12249 solver.cpp:229] Iteration 190500, loss = 2.14788 | |
| I0409 19:47:20.113804 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.5 | |
| I0409 19:47:20.113826 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 19:47:20.113839 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.8 | |
| I0409 19:47:20.113857 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.55986 (* 0.3 = 0.467959 loss) | |
| I0409 19:47:20.113873 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.468446 (* 0.3 = 0.140534 loss) | |
| I0409 19:47:20.113885 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.72 | |
| I0409 19:47:20.113898 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.920455 | |
| I0409 19:47:20.113909 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.96 | |
| I0409 19:47:20.113924 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.883922 (* 0.3 = 0.265177 loss) | |
| I0409 19:47:20.113939 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.256591 (* 0.3 = 0.0769775 loss) | |
| I0409 19:47:20.113951 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 1 | |
| I0409 19:47:20.113963 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.994318 | |
| I0409 19:47:20.113975 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 19:47:20.113988 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.0514849 (* 1 = 0.0514849 loss) | |
| I0409 19:47:20.114003 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0232809 (* 1 = 0.0232809 loss) | |
| I0409 19:47:20.114015 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 19:47:20.114028 12249 solver.cpp:245] Train net output #16: total_confidence = 0.682198 | |
| I0409 19:47:20.114042 12249 sgd_solver.cpp:106] Iteration 190500, lr = 0.00727857 | |
| I0409 19:52:53.490232 12249 solver.cpp:229] Iteration 191000, loss = 2.12955 | |
| I0409 19:52:53.490512 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.375 | |
| I0409 19:52:53.490533 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 19:52:53.490547 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.75 | |
| I0409 19:52:53.490563 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.83455 (* 0.3 = 0.550365 loss) | |
| I0409 19:52:53.490578 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.502919 (* 0.3 = 0.150876 loss) | |
| I0409 19:52:53.490592 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.7 | |
| I0409 19:52:53.490603 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 19:52:53.490615 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.875 | |
| I0409 19:52:53.490629 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.31936 (* 0.3 = 0.395807 loss) | |
| I0409 19:52:53.490643 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.383901 (* 0.3 = 0.11517 loss) | |
| I0409 19:52:53.490655 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.825 | |
| I0409 19:52:53.490669 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 19:52:53.490680 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.925 | |
| I0409 19:52:53.490694 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.840484 (* 1 = 0.840484 loss) | |
| I0409 19:52:53.490708 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.198389 (* 1 = 0.198389 loss) | |
| I0409 19:52:53.490721 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 19:52:53.490733 12249 solver.cpp:245] Train net output #16: total_confidence = 0.568223 | |
| I0409 19:52:53.490747 12249 sgd_solver.cpp:106] Iteration 191000, lr = 0.00727143 | |
| I0409 19:58:26.854962 12249 solver.cpp:229] Iteration 191500, loss = 2.214 | |
| I0409 19:58:26.855108 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.44898 | |
| I0409 19:58:26.855139 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 19:58:26.855165 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.714286 | |
| I0409 19:58:26.855190 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.01248 (* 0.3 = 0.603745 loss) | |
| I0409 19:58:26.855206 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.610413 (* 0.3 = 0.183124 loss) | |
| I0409 19:58:26.855218 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.55102 | |
| I0409 19:58:26.855231 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 19:58:26.855242 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.77551 | |
| I0409 19:58:26.855257 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.75666 (* 0.3 = 0.526997 loss) | |
| I0409 19:58:26.855271 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.525432 (* 0.3 = 0.15763 loss) | |
| I0409 19:58:26.855283 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.714286 | |
| I0409 19:58:26.855296 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.914773 | |
| I0409 19:58:26.855309 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.836735 | |
| I0409 19:58:26.855321 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.39544 (* 1 = 1.39544 loss) | |
| I0409 19:58:26.855336 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.401433 (* 1 = 0.401433 loss) | |
| I0409 19:58:26.855348 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 19:58:26.855360 12249 solver.cpp:245] Train net output #16: total_confidence = 0.352553 | |
| I0409 19:58:26.855375 12249 sgd_solver.cpp:106] Iteration 191500, lr = 0.00726429 | |
| I0409 20:04:00.224781 12249 solver.cpp:229] Iteration 192000, loss = 2.12901 | |
| I0409 20:04:00.225072 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.44898 | |
| I0409 20:04:00.225091 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 20:04:00.225105 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.755102 | |
| I0409 20:04:00.225121 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.66732 (* 0.3 = 0.500195 loss) | |
| I0409 20:04:00.225136 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.512352 (* 0.3 = 0.153706 loss) | |
| I0409 20:04:00.225149 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.632653 | |
| I0409 20:04:00.225162 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 20:04:00.225173 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.959184 | |
| I0409 20:04:00.225188 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.08187 (* 0.3 = 0.32456 loss) | |
| I0409 20:04:00.225203 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.316915 (* 0.3 = 0.0950746 loss) | |
| I0409 20:04:00.225214 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.979592 | |
| I0409 20:04:00.225227 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 20:04:00.225239 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 20:04:00.225253 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.117364 (* 1 = 0.117364 loss) | |
| I0409 20:04:00.225268 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0397128 (* 1 = 0.0397128 loss) | |
| I0409 20:04:00.225281 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 20:04:00.225293 12249 solver.cpp:245] Train net output #16: total_confidence = 0.571676 | |
| I0409 20:04:00.225307 12249 sgd_solver.cpp:106] Iteration 192000, lr = 0.00725714 | |
| I0409 20:09:33.594239 12249 solver.cpp:229] Iteration 192500, loss = 2.12022 | |
| I0409 20:09:33.594400 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.325581 | |
| I0409 20:09:33.594421 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 20:09:33.594434 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.627907 | |
| I0409 20:09:33.594452 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.45641 (* 0.3 = 0.736923 loss) | |
| I0409 20:09:33.594467 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.772989 (* 0.3 = 0.231897 loss) | |
| I0409 20:09:33.594480 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.581395 | |
| I0409 20:09:33.594492 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 20:09:33.594506 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.883721 | |
| I0409 20:09:33.594519 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.25935 (* 0.3 = 0.377806 loss) | |
| I0409 20:09:33.594533 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.402985 (* 0.3 = 0.120895 loss) | |
| I0409 20:09:33.594545 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.883721 | |
| I0409 20:09:33.594558 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 20:09:33.594570 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.930233 | |
| I0409 20:09:33.594585 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.42316 (* 1 = 0.42316 loss) | |
| I0409 20:09:33.594599 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.117474 (* 1 = 0.117474 loss) | |
| I0409 20:09:33.594614 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 20:09:33.594625 12249 solver.cpp:245] Train net output #16: total_confidence = 0.487467 | |
| I0409 20:09:33.594640 12249 sgd_solver.cpp:106] Iteration 192500, lr = 0.00725 | |
| I0409 20:15:06.970275 12249 solver.cpp:229] Iteration 193000, loss = 2.10378 | |
| I0409 20:15:06.970604 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 20:15:06.970626 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 20:15:06.970639 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.72 | |
| I0409 20:15:06.970655 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.72534 (* 0.3 = 0.517603 loss) | |
| I0409 20:15:06.970671 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.512056 (* 0.3 = 0.153617 loss) | |
| I0409 20:15:06.970685 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.66 | |
| I0409 20:15:06.970696 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 20:15:06.970708 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.92 | |
| I0409 20:15:06.970722 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.12163 (* 0.3 = 0.336491 loss) | |
| I0409 20:15:06.970737 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.343269 (* 0.3 = 0.102981 loss) | |
| I0409 20:15:06.970749 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.98 | |
| I0409 20:15:06.970762 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 20:15:06.970773 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 20:15:06.970788 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.145391 (* 1 = 0.145391 loss) | |
| I0409 20:15:06.970801 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0513221 (* 1 = 0.0513221 loss) | |
| I0409 20:15:06.970813 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 20:15:06.970825 12249 solver.cpp:245] Train net output #16: total_confidence = 0.463481 | |
| I0409 20:15:06.970840 12249 sgd_solver.cpp:106] Iteration 193000, lr = 0.00724286 | |
| I0409 20:20:40.336730 12249 solver.cpp:229] Iteration 193500, loss = 2.20543 | |
| I0409 20:20:40.336881 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.423077 | |
| I0409 20:20:40.336902 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.829545 | |
| I0409 20:20:40.336916 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.711538 | |
| I0409 20:20:40.336932 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.8602 (* 0.3 = 0.558059 loss) | |
| I0409 20:20:40.336947 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.576622 (* 0.3 = 0.172987 loss) | |
| I0409 20:20:40.336961 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.576923 | |
| I0409 20:20:40.336972 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 20:20:40.336984 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.903846 | |
| I0409 20:20:40.336998 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.27622 (* 0.3 = 0.382865 loss) | |
| I0409 20:20:40.337013 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.390662 (* 0.3 = 0.117199 loss) | |
| I0409 20:20:40.337025 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.923077 | |
| I0409 20:20:40.337038 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 20:20:40.337050 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 20:20:40.337064 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.253347 (* 1 = 0.253347 loss) | |
| I0409 20:20:40.337079 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0793728 (* 1 = 0.0793728 loss) | |
| I0409 20:20:40.337091 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 20:20:40.337103 12249 solver.cpp:245] Train net output #16: total_confidence = 0.363808 | |
| I0409 20:20:40.337118 12249 sgd_solver.cpp:106] Iteration 193500, lr = 0.00723571 | |
| I0409 20:26:13.712254 12249 solver.cpp:229] Iteration 194000, loss = 2.21969 | |
| I0409 20:26:13.712537 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 20:26:13.712555 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 20:26:13.712568 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.608696 | |
| I0409 20:26:13.712585 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.92352 (* 0.3 = 0.577056 loss) | |
| I0409 20:26:13.712600 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.549597 (* 0.3 = 0.164879 loss) | |
| I0409 20:26:13.712613 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.652174 | |
| I0409 20:26:13.712625 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 20:26:13.712638 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.869565 | |
| I0409 20:26:13.712653 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.11489 (* 0.3 = 0.334468 loss) | |
| I0409 20:26:13.712668 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.316887 (* 0.3 = 0.095066 loss) | |
| I0409 20:26:13.712680 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.956522 | |
| I0409 20:26:13.712692 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 20:26:13.712704 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.956522 | |
| I0409 20:26:13.712719 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.258016 (* 1 = 0.258016 loss) | |
| I0409 20:26:13.712733 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0728882 (* 1 = 0.0728882 loss) | |
| I0409 20:26:13.712746 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 20:26:13.712759 12249 solver.cpp:245] Train net output #16: total_confidence = 0.505175 | |
| I0409 20:26:13.712774 12249 sgd_solver.cpp:106] Iteration 194000, lr = 0.00722857 | |
| I0409 20:31:47.092473 12249 solver.cpp:229] Iteration 194500, loss = 2.11596 | |
| I0409 20:31:47.092628 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.367347 | |
| I0409 20:31:47.092648 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.801136 | |
| I0409 20:31:47.092663 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.510204 | |
| I0409 20:31:47.092679 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.2178 (* 0.3 = 0.665339 loss) | |
| I0409 20:31:47.092694 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.665192 (* 0.3 = 0.199558 loss) | |
| I0409 20:31:47.092708 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.489796 | |
| I0409 20:31:47.092720 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.846591 | |
| I0409 20:31:47.092732 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.795918 | |
| I0409 20:31:47.092749 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.54123 (* 0.3 = 0.46237 loss) | |
| I0409 20:31:47.092764 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.473401 (* 0.3 = 0.14202 loss) | |
| I0409 20:31:47.092777 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.632653 | |
| I0409 20:31:47.092789 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.897727 | |
| I0409 20:31:47.092802 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.857143 | |
| I0409 20:31:47.092815 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.01301 (* 1 = 1.01301 loss) | |
| I0409 20:31:47.092829 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.284122 (* 1 = 0.284122 loss) | |
| I0409 20:31:47.092842 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 20:31:47.092854 12249 solver.cpp:245] Train net output #16: total_confidence = 0.359319 | |
| I0409 20:31:47.092869 12249 sgd_solver.cpp:106] Iteration 194500, lr = 0.00722143 | |
| I0409 20:37:20.062177 12249 solver.cpp:338] Iteration 195000, Testing net (#0) | |
| I0409 20:38:01.248199 12249 solver.cpp:393] Test loss: 1.93431 | |
| I0409 20:38:01.248292 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.533626 | |
| I0409 20:38:01.248318 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.879867 | |
| I0409 20:38:01.248340 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.793777 | |
| I0409 20:38:01.248366 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.63645 (* 0.3 = 0.490934 loss) | |
| I0409 20:38:01.248391 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.431291 (* 0.3 = 0.129387 loss) | |
| I0409 20:38:01.248412 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.695292 | |
| I0409 20:38:01.248435 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.92041 | |
| I0409 20:38:01.248456 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.89191 | |
| I0409 20:38:01.248497 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.09699 (* 0.3 = 0.329098 loss) | |
| I0409 20:38:01.248527 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.294744 (* 0.3 = 0.0884231 loss) | |
| I0409 20:38:01.248548 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.829782 | |
| I0409 20:38:01.248569 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.958637 | |
| I0409 20:38:01.248589 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.911781 | |
| I0409 20:38:01.248613 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.712897 (* 1 = 0.712897 loss) | |
| I0409 20:38:01.248639 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.183568 (* 1 = 0.183568 loss) | |
| I0409 20:38:01.248661 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.601 | |
| I0409 20:38:01.248680 12249 solver.cpp:406] Test net output #16: total_confidence = 0.541839 | |
| I0409 20:38:01.624584 12249 solver.cpp:229] Iteration 195000, loss = 2.15499 | |
| I0409 20:38:01.624642 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.34 | |
| I0409 20:38:01.624660 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 20:38:01.624675 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.7 | |
| I0409 20:38:01.624691 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.11268 (* 0.3 = 0.633803 loss) | |
| I0409 20:38:01.624706 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.648719 (* 0.3 = 0.194616 loss) | |
| I0409 20:38:01.624718 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.54 | |
| I0409 20:38:01.624732 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 20:38:01.624743 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.88 | |
| I0409 20:38:01.624758 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.27413 (* 0.3 = 0.38224 loss) | |
| I0409 20:38:01.624773 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.403961 (* 0.3 = 0.121188 loss) | |
| I0409 20:38:01.624784 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.82 | |
| I0409 20:38:01.624796 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.943182 | |
| I0409 20:38:01.624809 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.92 | |
| I0409 20:38:01.624822 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.727543 (* 1 = 0.727543 loss) | |
| I0409 20:38:01.624840 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.23008 (* 1 = 0.23008 loss) | |
| I0409 20:38:01.624853 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 20:38:01.624866 12249 solver.cpp:245] Train net output #16: total_confidence = 0.329482 | |
| I0409 20:38:01.624881 12249 sgd_solver.cpp:106] Iteration 195000, lr = 0.00721429 | |
| I0409 20:43:34.839449 12249 solver.cpp:229] Iteration 195500, loss = 2.14795 | |
| I0409 20:43:34.839747 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.314815 | |
| I0409 20:43:34.839768 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 20:43:34.839781 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.592593 | |
| I0409 20:43:34.839797 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.25462 (* 0.3 = 0.676387 loss) | |
| I0409 20:43:34.839812 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.714273 (* 0.3 = 0.214282 loss) | |
| I0409 20:43:34.839825 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.444444 | |
| I0409 20:43:34.839838 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.829545 | |
| I0409 20:43:34.839850 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.740741 | |
| I0409 20:43:34.839864 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.79451 (* 0.3 = 0.538354 loss) | |
| I0409 20:43:34.839879 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.562527 (* 0.3 = 0.168758 loss) | |
| I0409 20:43:34.839890 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.833333 | |
| I0409 20:43:34.839902 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 20:43:34.839913 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.907407 | |
| I0409 20:43:34.839928 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.741703 (* 1 = 0.741703 loss) | |
| I0409 20:43:34.839942 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.240222 (* 1 = 0.240222 loss) | |
| I0409 20:43:34.839954 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 20:43:34.839967 12249 solver.cpp:245] Train net output #16: total_confidence = 0.26995 | |
| I0409 20:43:34.839982 12249 sgd_solver.cpp:106] Iteration 195500, lr = 0.00720714 | |
| I0409 20:49:08.212800 12249 solver.cpp:229] Iteration 196000, loss = 2.18081 | |
| I0409 20:49:08.212940 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.45 | |
| I0409 20:49:08.212970 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 20:49:08.212996 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.625 | |
| I0409 20:49:08.213018 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.91799 (* 0.3 = 0.575398 loss) | |
| I0409 20:49:08.213034 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.497793 (* 0.3 = 0.149338 loss) | |
| I0409 20:49:08.213047 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.6 | |
| I0409 20:49:08.213059 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 20:49:08.213071 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.825 | |
| I0409 20:49:08.213085 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.12594 (* 0.3 = 0.337783 loss) | |
| I0409 20:49:08.213100 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.295729 (* 0.3 = 0.0887187 loss) | |
| I0409 20:49:08.213112 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.875 | |
| I0409 20:49:08.213125 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 20:49:08.213136 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.975 | |
| I0409 20:49:08.213150 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.440383 (* 1 = 0.440383 loss) | |
| I0409 20:49:08.213165 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.105581 (* 1 = 0.105581 loss) | |
| I0409 20:49:08.213177 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 20:49:08.213189 12249 solver.cpp:245] Train net output #16: total_confidence = 0.420241 | |
| I0409 20:49:08.213204 12249 sgd_solver.cpp:106] Iteration 196000, lr = 0.0072 | |
| I0409 20:54:41.564728 12249 solver.cpp:229] Iteration 196500, loss = 2.19159 | |
| I0409 20:54:41.564988 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.444444 | |
| I0409 20:54:41.565007 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 20:54:41.565021 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.711111 | |
| I0409 20:54:41.565037 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.80008 (* 0.3 = 0.540024 loss) | |
| I0409 20:54:41.565052 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.511394 (* 0.3 = 0.153418 loss) | |
| I0409 20:54:41.565064 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.511111 | |
| I0409 20:54:41.565076 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.852273 | |
| I0409 20:54:41.565088 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.866667 | |
| I0409 20:54:41.565101 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.35561 (* 0.3 = 0.406682 loss) | |
| I0409 20:54:41.565116 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.422546 (* 0.3 = 0.126764 loss) | |
| I0409 20:54:41.565129 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.955556 | |
| I0409 20:54:41.565140 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 20:54:41.565152 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.955556 | |
| I0409 20:54:41.565166 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.305615 (* 1 = 0.305615 loss) | |
| I0409 20:54:41.565186 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.117045 (* 1 = 0.117045 loss) | |
| I0409 20:54:41.565198 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.875 | |
| I0409 20:54:41.565210 12249 solver.cpp:245] Train net output #16: total_confidence = 0.604933 | |
| I0409 20:54:41.565224 12249 sgd_solver.cpp:106] Iteration 196500, lr = 0.00719286 | |
| I0409 21:00:14.941545 12249 solver.cpp:229] Iteration 197000, loss = 2.14113 | |
| I0409 21:00:14.941712 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.365854 | |
| I0409 21:00:14.941737 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 21:00:14.941752 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.634146 | |
| I0409 21:00:14.941769 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.05473 (* 0.3 = 0.616418 loss) | |
| I0409 21:00:14.941784 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.629336 (* 0.3 = 0.188801 loss) | |
| I0409 21:00:14.941797 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.560976 | |
| I0409 21:00:14.941809 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 21:00:14.941822 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.829268 | |
| I0409 21:00:14.941835 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.28905 (* 0.3 = 0.386715 loss) | |
| I0409 21:00:14.941850 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.474474 (* 0.3 = 0.142342 loss) | |
| I0409 21:00:14.941864 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.926829 | |
| I0409 21:00:14.941875 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.960227 | |
| I0409 21:00:14.941887 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.97561 | |
| I0409 21:00:14.941901 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.23649 (* 1 = 0.23649 loss) | |
| I0409 21:00:14.941918 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.204433 (* 1 = 0.204433 loss) | |
| I0409 21:00:14.941931 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 21:00:14.941943 12249 solver.cpp:245] Train net output #16: total_confidence = 0.626897 | |
| I0409 21:00:14.941958 12249 sgd_solver.cpp:106] Iteration 197000, lr = 0.00718571 | |
| I0409 21:05:48.305181 12249 solver.cpp:229] Iteration 197500, loss = 2.18545 | |
| I0409 21:05:48.305465 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.313726 | |
| I0409 21:05:48.305485 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 21:05:48.305498 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.54902 | |
| I0409 21:05:48.305516 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.24069 (* 0.3 = 0.672206 loss) | |
| I0409 21:05:48.305531 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.697598 (* 0.3 = 0.209279 loss) | |
| I0409 21:05:48.305543 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.54902 | |
| I0409 21:05:48.305557 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 21:05:48.305568 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.784314 | |
| I0409 21:05:48.305583 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.53146 (* 0.3 = 0.459437 loss) | |
| I0409 21:05:48.305596 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.481727 (* 0.3 = 0.144518 loss) | |
| I0409 21:05:48.305608 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.882353 | |
| I0409 21:05:48.305621 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.965909 | |
| I0409 21:05:48.305634 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.941176 | |
| I0409 21:05:48.305649 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.549704 (* 1 = 0.549704 loss) | |
| I0409 21:05:48.305662 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.168926 (* 1 = 0.168926 loss) | |
| I0409 21:05:48.305675 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 21:05:48.305687 12249 solver.cpp:245] Train net output #16: total_confidence = 0.546465 | |
| I0409 21:05:48.305701 12249 sgd_solver.cpp:106] Iteration 197500, lr = 0.00717857 | |
| I0409 21:11:21.678614 12249 solver.cpp:229] Iteration 198000, loss = 2.13253 | |
| I0409 21:11:21.678724 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.285714 | |
| I0409 21:11:21.678753 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 21:11:21.678774 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.510204 | |
| I0409 21:11:21.678792 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.37923 (* 0.3 = 0.71377 loss) | |
| I0409 21:11:21.678807 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.716425 (* 0.3 = 0.214928 loss) | |
| I0409 21:11:21.678820 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.612245 | |
| I0409 21:11:21.678833 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 21:11:21.678845 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.877551 | |
| I0409 21:11:21.678858 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.30253 (* 0.3 = 0.39076 loss) | |
| I0409 21:11:21.678874 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.408825 (* 0.3 = 0.122647 loss) | |
| I0409 21:11:21.678885 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.938776 | |
| I0409 21:11:21.678897 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 21:11:21.678910 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 21:11:21.678925 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.211945 (* 1 = 0.211945 loss) | |
| I0409 21:11:21.678938 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0714908 (* 1 = 0.0714908 loss) | |
| I0409 21:11:21.678951 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.625 | |
| I0409 21:11:21.678963 12249 solver.cpp:245] Train net output #16: total_confidence = 0.377464 | |
| I0409 21:11:21.678977 12249 sgd_solver.cpp:106] Iteration 198000, lr = 0.00717143 | |
| I0409 21:16:55.047655 12249 solver.cpp:229] Iteration 198500, loss = 2.16332 | |
| I0409 21:16:55.047889 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.276596 | |
| I0409 21:16:55.047909 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.795455 | |
| I0409 21:16:55.047921 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.617021 | |
| I0409 21:16:55.047937 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.09185 (* 0.3 = 0.627555 loss) | |
| I0409 21:16:55.047952 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.625587 (* 0.3 = 0.187676 loss) | |
| I0409 21:16:55.047965 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.574468 | |
| I0409 21:16:55.047976 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 21:16:55.047988 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.787234 | |
| I0409 21:16:55.048002 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.6845 (* 0.3 = 0.50535 loss) | |
| I0409 21:16:55.048017 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.494737 (* 0.3 = 0.148421 loss) | |
| I0409 21:16:55.048029 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.723404 | |
| I0409 21:16:55.048041 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 21:16:55.048053 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.87234 | |
| I0409 21:16:55.048068 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.10967 (* 1 = 1.10967 loss) | |
| I0409 21:16:55.048081 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.316969 (* 1 = 0.316969 loss) | |
| I0409 21:16:55.048094 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 21:16:55.048105 12249 solver.cpp:245] Train net output #16: total_confidence = 0.31901 | |
| I0409 21:16:55.048120 12249 sgd_solver.cpp:106] Iteration 198500, lr = 0.00716429 | |
| I0409 21:22:28.411051 12249 solver.cpp:229] Iteration 199000, loss = 2.14067 | |
| I0409 21:22:28.411396 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.456522 | |
| I0409 21:22:28.411418 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 21:22:28.411432 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.73913 | |
| I0409 21:22:28.411448 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.75909 (* 0.3 = 0.527727 loss) | |
| I0409 21:22:28.411463 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.495582 (* 0.3 = 0.148675 loss) | |
| I0409 21:22:28.411476 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.652174 | |
| I0409 21:22:28.411489 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 21:22:28.411501 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.934783 | |
| I0409 21:22:28.411515 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.11863 (* 0.3 = 0.335588 loss) | |
| I0409 21:22:28.411530 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.317603 (* 0.3 = 0.0952808 loss) | |
| I0409 21:22:28.411542 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.956522 | |
| I0409 21:22:28.411555 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.988636 | |
| I0409 21:22:28.411566 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 21:22:28.411581 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.164696 (* 1 = 0.164696 loss) | |
| I0409 21:22:28.411594 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.0496942 (* 1 = 0.0496942 loss) | |
| I0409 21:22:28.411607 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 21:22:28.411619 12249 solver.cpp:245] Train net output #16: total_confidence = 0.504681 | |
| I0409 21:22:28.411633 12249 sgd_solver.cpp:106] Iteration 199000, lr = 0.00715714 | |
| I0409 21:28:01.784289 12249 solver.cpp:229] Iteration 199500, loss = 2.16072 | |
| I0409 21:28:01.784405 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.4 | |
| I0409 21:28:01.784425 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.818182 | |
| I0409 21:28:01.784438 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.66 | |
| I0409 21:28:01.784454 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.0427 (* 0.3 = 0.612811 loss) | |
| I0409 21:28:01.784469 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.620475 (* 0.3 = 0.186143 loss) | |
| I0409 21:28:01.784482 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.7 | |
| I0409 21:28:01.784494 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.897727 | |
| I0409 21:28:01.784507 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.84 | |
| I0409 21:28:01.784521 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.29122 (* 0.3 = 0.387365 loss) | |
| I0409 21:28:01.784550 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.426139 (* 0.3 = 0.127842 loss) | |
| I0409 21:28:01.784564 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.84 | |
| I0409 21:28:01.784576 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 21:28:01.784589 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.96 | |
| I0409 21:28:01.784602 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.542244 (* 1 = 0.542244 loss) | |
| I0409 21:28:01.784616 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.180759 (* 1 = 0.180759 loss) | |
| I0409 21:28:01.784628 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.375 | |
| I0409 21:28:01.784642 12249 solver.cpp:245] Train net output #16: total_confidence = 0.348674 | |
| I0409 21:28:01.784657 12249 sgd_solver.cpp:106] Iteration 199500, lr = 0.00715 | |
| I0409 21:33:34.777505 12249 solver.cpp:456] Snapshotting to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_200000.caffemodel | |
| I0409 21:33:35.385962 12249 sgd_solver.cpp:273] Snapshotting solver state to binary proto file /mnt/snapshots/mixed_lstm10_bn_iter_200000.solverstate | |
| I0409 21:33:35.629808 12249 solver.cpp:338] Iteration 200000, Testing net (#0) | |
| I0409 21:34:17.001790 12249 solver.cpp:393] Test loss: 1.88013 | |
| I0409 21:34:17.001938 12249 solver.cpp:406] Test net output #0: loss1/accuracy = 0.527624 | |
| I0409 21:34:17.001972 12249 solver.cpp:406] Test net output #1: loss1/accuracy_incl_empty = 0.878685 | |
| I0409 21:34:17.002001 12249 solver.cpp:406] Test net output #2: loss1/accuracy_top3 = 0.801222 | |
| I0409 21:34:17.002033 12249 solver.cpp:406] Test net output #3: loss1/cross_entropy_loss = 1.58444 (* 0.3 = 0.475332 loss) | |
| I0409 21:34:17.002065 12249 solver.cpp:406] Test net output #4: loss1/cross_entropy_loss_incl_empty = 0.417491 (* 0.3 = 0.125247 loss) | |
| I0409 21:34:17.002091 12249 solver.cpp:406] Test net output #5: loss2/accuracy = 0.716128 | |
| I0409 21:34:17.002116 12249 solver.cpp:406] Test net output #6: loss2/accuracy_incl_empty = 0.926138 | |
| I0409 21:34:17.002141 12249 solver.cpp:406] Test net output #7: loss2/accuracy_top3 = 0.900219 | |
| I0409 21:34:17.002169 12249 solver.cpp:406] Test net output #8: loss2/cross_entropy_loss = 1.04778 (* 0.3 = 0.314334 loss) | |
| I0409 21:34:17.002192 12249 solver.cpp:406] Test net output #9: loss2/cross_entropy_loss_incl_empty = 0.276314 (* 0.3 = 0.0828941 loss) | |
| I0409 21:34:17.002214 12249 solver.cpp:406] Test net output #10: loss3/accuracy = 0.839961 | |
| I0409 21:34:17.002235 12249 solver.cpp:406] Test net output #11: loss3/accuracy_incl_empty = 0.960819 | |
| I0409 21:34:17.002256 12249 solver.cpp:406] Test net output #12: loss3/accuracy_top3 = 0.923835 | |
| I0409 21:34:17.002282 12249 solver.cpp:406] Test net output #13: loss3/cross_entropy_loss = 0.700885 (* 1 = 0.700885 loss) | |
| I0409 21:34:17.002306 12249 solver.cpp:406] Test net output #14: loss3/cross_entropy_loss_incl_empty = 0.181434 (* 1 = 0.181434 loss) | |
| I0409 21:34:17.002326 12249 solver.cpp:406] Test net output #15: total_accuracy = 0.618 | |
| I0409 21:34:17.002344 12249 solver.cpp:406] Test net output #16: total_confidence = 0.589672 | |
| I0409 21:34:17.376088 12249 solver.cpp:229] Iteration 200000, loss = 2.12325 | |
| I0409 21:34:17.376147 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.434783 | |
| I0409 21:34:17.376165 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.823864 | |
| I0409 21:34:17.376178 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.717391 | |
| I0409 21:34:17.376194 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.89517 (* 0.3 = 0.56855 loss) | |
| I0409 21:34:17.376209 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.562809 (* 0.3 = 0.168843 loss) | |
| I0409 21:34:17.376225 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.478261 | |
| I0409 21:34:17.376240 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.857955 | |
| I0409 21:34:17.376251 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.717391 | |
| I0409 21:34:17.376266 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 2.00524 (* 0.3 = 0.601573 loss) | |
| I0409 21:34:17.376281 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.548116 (* 0.3 = 0.164435 loss) | |
| I0409 21:34:17.376293 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.717391 | |
| I0409 21:34:17.376307 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.926136 | |
| I0409 21:34:17.376318 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.891304 | |
| I0409 21:34:17.376332 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 1.14992 (* 1 = 1.14992 loss) | |
| I0409 21:34:17.376348 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.308549 (* 1 = 0.308549 loss) | |
| I0409 21:34:17.376359 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 21:34:17.376373 12249 solver.cpp:245] Train net output #16: total_confidence = 0.575486 | |
| I0409 21:34:17.376389 12249 sgd_solver.cpp:106] Iteration 200000, lr = 0.00714286 | |
| I0409 21:39:50.871021 12249 solver.cpp:229] Iteration 200500, loss = 2.13235 | |
| I0409 21:39:50.871186 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.291667 | |
| I0409 21:39:50.871211 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.789773 | |
| I0409 21:39:50.871224 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.5625 | |
| I0409 21:39:50.871242 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.43944 (* 0.3 = 0.731833 loss) | |
| I0409 21:39:50.871258 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.737243 (* 0.3 = 0.221173 loss) | |
| I0409 21:39:50.871270 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.520833 | |
| I0409 21:39:50.871282 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.863636 | |
| I0409 21:39:50.871294 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.770833 | |
| I0409 21:39:50.871309 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.63912 (* 0.3 = 0.491736 loss) | |
| I0409 21:39:50.871325 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.463689 (* 0.3 = 0.139107 loss) | |
| I0409 21:39:50.871337 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.916667 | |
| I0409 21:39:50.871350 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 21:39:50.871361 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 1 | |
| I0409 21:39:50.871376 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.262327 (* 1 = 0.262327 loss) | |
| I0409 21:39:50.871390 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.086173 (* 1 = 0.086173 loss) | |
| I0409 21:39:50.871402 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 21:39:50.871415 12249 solver.cpp:245] Train net output #16: total_confidence = 0.391667 | |
| I0409 21:39:50.871429 12249 sgd_solver.cpp:106] Iteration 200500, lr = 0.00713571 | |
| I0409 21:45:24.235496 12249 solver.cpp:229] Iteration 201000, loss = 2.16955 | |
| I0409 21:45:24.235772 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.319149 | |
| I0409 21:45:24.235792 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.8125 | |
| I0409 21:45:24.235806 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.553191 | |
| I0409 21:45:24.235823 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.41789 (* 0.3 = 0.725366 loss) | |
| I0409 21:45:24.235838 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.685641 (* 0.3 = 0.205692 loss) | |
| I0409 21:45:24.235852 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.553191 | |
| I0409 21:45:24.235864 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.875 | |
| I0409 21:45:24.235877 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.808511 | |
| I0409 21:45:24.235891 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.6101 (* 0.3 = 0.483029 loss) | |
| I0409 21:45:24.235905 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.486166 (* 0.3 = 0.14585 loss) | |
| I0409 21:45:24.235918 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.829787 | |
| I0409 21:45:24.235930 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.948864 | |
| I0409 21:45:24.235944 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.957447 | |
| I0409 21:45:24.235957 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.56874 (* 1 = 0.56874 loss) | |
| I0409 21:45:24.235972 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.175889 (* 1 = 0.175889 loss) | |
| I0409 21:45:24.235985 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 21:45:24.235996 12249 solver.cpp:245] Train net output #16: total_confidence = 0.304581 | |
| I0409 21:45:24.236013 12249 sgd_solver.cpp:106] Iteration 201000, lr = 0.00712857 | |
| I0409 21:50:57.612284 12249 solver.cpp:229] Iteration 201500, loss = 2.14689 | |
| I0409 21:50:57.612473 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.404762 | |
| I0409 21:50:57.612493 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 21:50:57.612507 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.690476 | |
| I0409 21:50:57.612524 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.19601 (* 0.3 = 0.658804 loss) | |
| I0409 21:50:57.612540 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.616052 (* 0.3 = 0.184816 loss) | |
| I0409 21:50:57.612552 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.452381 | |
| I0409 21:50:57.612565 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.840909 | |
| I0409 21:50:57.612578 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.833333 | |
| I0409 21:50:57.612592 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.61044 (* 0.3 = 0.483132 loss) | |
| I0409 21:50:57.612607 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.451649 (* 0.3 = 0.135495 loss) | |
| I0409 21:50:57.612619 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.833333 | |
| I0409 21:50:57.612632 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.954545 | |
| I0409 21:50:57.612643 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.97619 | |
| I0409 21:50:57.612658 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.4557 (* 1 = 0.4557 loss) | |
| I0409 21:50:57.612673 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.132018 (* 1 = 0.132018 loss) | |
| I0409 21:50:57.612685 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 21:50:57.612697 12249 solver.cpp:245] Train net output #16: total_confidence = 0.474416 | |
| I0409 21:50:57.612712 12249 sgd_solver.cpp:106] Iteration 201500, lr = 0.00712143 | |
| I0409 21:56:30.991586 12249 solver.cpp:229] Iteration 202000, loss = 2.10746 | |
| I0409 21:56:30.991925 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.509804 | |
| I0409 21:56:30.991946 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.852273 | |
| I0409 21:56:30.991960 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.764706 | |
| I0409 21:56:30.991976 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.51007 (* 0.3 = 0.45302 loss) | |
| I0409 21:56:30.991991 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.458439 (* 0.3 = 0.137532 loss) | |
| I0409 21:56:30.992004 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.686275 | |
| I0409 21:56:30.992017 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 21:56:30.992028 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.882353 | |
| I0409 21:56:30.992043 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 0.935842 (* 0.3 = 0.280752 loss) | |
| I0409 21:56:30.992058 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.279899 (* 0.3 = 0.0839696 loss) | |
| I0409 21:56:30.992070 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.901961 | |
| I0409 21:56:30.992082 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.971591 | |
| I0409 21:56:30.992094 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.941176 | |
| I0409 21:56:30.992110 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.387519 (* 1 = 0.387519 loss) | |
| I0409 21:56:30.992123 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.11835 (* 1 = 0.11835 loss) | |
| I0409 21:56:30.992136 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.75 | |
| I0409 21:56:30.992147 12249 solver.cpp:245] Train net output #16: total_confidence = 0.515193 | |
| I0409 21:56:30.992162 12249 sgd_solver.cpp:106] Iteration 202000, lr = 0.00711429 | |
| I0409 22:02:04.686964 12249 solver.cpp:229] Iteration 202500, loss = 2.17605 | |
| I0409 22:02:04.687104 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.425532 | |
| I0409 22:02:04.687124 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.846591 | |
| I0409 22:02:04.687137 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.723404 | |
| I0409 22:02:04.687155 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.88349 (* 0.3 = 0.565048 loss) | |
| I0409 22:02:04.687170 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.518979 (* 0.3 = 0.155694 loss) | |
| I0409 22:02:04.687182 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.680851 | |
| I0409 22:02:04.687194 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.903409 | |
| I0409 22:02:04.687206 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.87234 | |
| I0409 22:02:04.687221 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.23751 (* 0.3 = 0.371253 loss) | |
| I0409 22:02:04.687234 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.354022 (* 0.3 = 0.106207 loss) | |
| I0409 22:02:04.687247 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.914894 | |
| I0409 22:02:04.687258 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.977273 | |
| I0409 22:02:04.687270 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.978723 | |
| I0409 22:02:04.687285 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.369778 (* 1 = 0.369778 loss) | |
| I0409 22:02:04.687299 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.101003 (* 1 = 0.101003 loss) | |
| I0409 22:02:04.687311 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.5 | |
| I0409 22:02:04.687324 12249 solver.cpp:245] Train net output #16: total_confidence = 0.375216 | |
| I0409 22:02:04.687337 12249 sgd_solver.cpp:106] Iteration 202500, lr = 0.00710714 | |
| I0409 22:07:38.069916 12249 solver.cpp:229] Iteration 203000, loss = 2.08689 | |
| I0409 22:07:38.070206 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.435897 | |
| I0409 22:07:38.070226 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.857955 | |
| I0409 22:07:38.070240 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.564103 | |
| I0409 22:07:38.070257 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 2.4531 (* 0.3 = 0.735931 loss) | |
| I0409 22:07:38.070272 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.603539 (* 0.3 = 0.181062 loss) | |
| I0409 22:07:38.070288 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.512821 | |
| I0409 22:07:38.070302 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.880682 | |
| I0409 22:07:38.070313 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.74359 | |
| I0409 22:07:38.070327 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.50224 (* 0.3 = 0.450671 loss) | |
| I0409 22:07:38.070343 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.392845 (* 0.3 = 0.117854 loss) | |
| I0409 22:07:38.070354 12249 solver.cpp:245] Train net output #10: loss3/accuracy = 0.74359 | |
| I0409 22:07:38.070368 12249 solver.cpp:245] Train net output #11: loss3/accuracy_incl_empty = 0.9375 | |
| I0409 22:07:38.070379 12249 solver.cpp:245] Train net output #12: loss3/accuracy_top3 = 0.820513 | |
| I0409 22:07:38.070394 12249 solver.cpp:245] Train net output #13: loss3/cross_entropy_loss = 0.827973 (* 1 = 0.827973 loss) | |
| I0409 22:07:38.070408 12249 solver.cpp:245] Train net output #14: loss3/cross_entropy_loss_incl_empty = 0.2063 (* 1 = 0.2063 loss) | |
| I0409 22:07:38.070420 12249 solver.cpp:245] Train net output #15: total_accuracy = 0.25 | |
| I0409 22:07:38.070432 12249 solver.cpp:245] Train net output #16: total_confidence = 0.421686 | |
| I0409 22:07:38.070448 12249 sgd_solver.cpp:106] Iteration 203000, lr = 0.0071 | |
| I0409 22:13:11.439563 12249 solver.cpp:229] Iteration 203500, loss = 2.07353 | |
| I0409 22:13:11.439934 12249 solver.cpp:245] Train net output #0: loss1/accuracy = 0.425532 | |
| I0409 22:13:11.439955 12249 solver.cpp:245] Train net output #1: loss1/accuracy_incl_empty = 0.835227 | |
| I0409 22:13:11.439970 12249 solver.cpp:245] Train net output #2: loss1/accuracy_top3 = 0.765957 | |
| I0409 22:13:11.439985 12249 solver.cpp:245] Train net output #3: loss1/cross_entropy_loss = 1.73913 (* 0.3 = 0.52174 loss) | |
| I0409 22:13:11.440001 12249 solver.cpp:245] Train net output #4: loss1/cross_entropy_loss_incl_empty = 0.508937 (* 0.3 = 0.152681 loss) | |
| I0409 22:13:11.440016 12249 solver.cpp:245] Train net output #5: loss2/accuracy = 0.680851 | |
| I0409 22:13:11.440027 12249 solver.cpp:245] Train net output #6: loss2/accuracy_incl_empty = 0.909091 | |
| I0409 22:13:11.440039 12249 solver.cpp:245] Train net output #7: loss2/accuracy_top3 = 0.851064 | |
| I0409 22:13:11.440054 12249 solver.cpp:245] Train net output #8: loss2/cross_entropy_loss = 1.14654 (* 0.3 = 0.343961 loss) | |
| I0409 22:13:11.440069 12249 solver.cpp:245] Train net output #9: loss2/cross_entropy_loss_incl_empty = 0.331854 (* 0.3 = 0.0 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment