yoshitomo-matsubara's picture
initial commit
b96d3f6
2021-06-01 19:42:00,212 INFO __main__ Namespace(adjust_lr=False, config='torchdistill/configs/sample/glue/qnli/kd/bert_base_uncased_from_bert_large_uncased.yaml', log='log/glue/qnli/kd/bert_base_uncased_from_bert_large_uncased.txt', private_output='leaderboard/glue/kd/bert_base_uncased_from_bert_large_uncased/', seed=None, student_only=False, task_name='qnli', test_only=False, world_size=1)
2021-06-01 19:42:00,284 INFO __main__ Distributed environment: NO
Num processes: 1
Process index: 0
Local process index: 0
Device: cuda
Use FP16 precision: True
2021-06-01 19:42:00,637 INFO filelock Lock 139934447891088 acquired on /root/.cache/huggingface/transformers/20af61132d3eccaf2a0b4fd9a18767272ba96e31e9a0b1d0035ef88dcdf1825b.92355a00abb9ce3df48e653f7303627944e38f408924eaed14015d4b5ab463c2.lock
2021-06-01 19:42:00,984 INFO filelock Lock 139934447891088 released on /root/.cache/huggingface/transformers/20af61132d3eccaf2a0b4fd9a18767272ba96e31e9a0b1d0035ef88dcdf1825b.92355a00abb9ce3df48e653f7303627944e38f408924eaed14015d4b5ab463c2.lock
2021-06-01 19:42:01,689 INFO filelock Lock 139934440344528 acquired on /root/.cache/huggingface/transformers/a89adc171199945c88c6380d6e359e2f2f1d9e33fbf2914840ccbcae2a88cc90.d789d64ebfe299b0e416afc4a169632f903f693095b4629a7ea271d5a0cf2c99.lock
2021-06-01 19:42:02,212 INFO filelock Lock 139934440344528 released on /root/.cache/huggingface/transformers/a89adc171199945c88c6380d6e359e2f2f1d9e33fbf2914840ccbcae2a88cc90.d789d64ebfe299b0e416afc4a169632f903f693095b4629a7ea271d5a0cf2c99.lock
2021-06-01 19:42:02,562 INFO filelock Lock 139934440344528 acquired on /root/.cache/huggingface/transformers/e40cf262b8ee3363b76e16e6b8d7a7cff091ce90d0df585990d20879a18b62d6.6dc9f54d5893dc361ac6ccee1865622847ad90bf0536eeb2043f3e3e2f41078a.lock
2021-06-01 19:42:03,100 INFO filelock Lock 139934440344528 released on /root/.cache/huggingface/transformers/e40cf262b8ee3363b76e16e6b8d7a7cff091ce90d0df585990d20879a18b62d6.6dc9f54d5893dc361ac6ccee1865622847ad90bf0536eeb2043f3e3e2f41078a.lock
2021-06-01 19:42:03,804 INFO filelock Lock 139934407962256 acquired on /root/.cache/huggingface/transformers/8af4ffda1cdd3bddb5fdcc1e67501f432ccc42db734ab6108c0013a2f653330e.dd8bd9bfd3664b530ea4e645105f557769387b3da9f79bdb55ed556bdd80611d.lock
2021-06-01 19:42:04,154 INFO filelock Lock 139934407962256 released on /root/.cache/huggingface/transformers/8af4ffda1cdd3bddb5fdcc1e67501f432ccc42db734ab6108c0013a2f653330e.dd8bd9bfd3664b530ea4e645105f557769387b3da9f79bdb55ed556bdd80611d.lock
2021-06-01 19:42:04,504 INFO filelock Lock 139934407988304 acquired on /root/.cache/huggingface/transformers/ceadc7427e9956e9cfd059a42937d2e2622d3da1f2c317b34b12dfc4a7d55d3b.0f95f2171d2c33a9e9e088c1e5decb2dfb3a22fb00d904f96183827da9540426.lock
2021-06-01 19:42:04,853 INFO filelock Lock 139934407988304 released on /root/.cache/huggingface/transformers/ceadc7427e9956e9cfd059a42937d2e2622d3da1f2c317b34b12dfc4a7d55d3b.0f95f2171d2c33a9e9e088c1e5decb2dfb3a22fb00d904f96183827da9540426.lock
2021-06-01 19:42:05,233 INFO filelock Lock 139934407943440 acquired on /root/.cache/huggingface/transformers/e9885efd6faed99edb75942a797fb277852d1c5e496344391594af18bdb6aa75.142b7305af89599e15e0af127ab907dbd903a36257191ae84062060491593a1e.lock
2021-06-01 19:42:28,226 INFO filelock Lock 139934407943440 released on /root/.cache/huggingface/transformers/e9885efd6faed99edb75942a797fb277852d1c5e496344391594af18bdb6aa75.142b7305af89599e15e0af127ab907dbd903a36257191ae84062060491593a1e.lock
2021-06-01 19:42:32,093 INFO filelock Lock 139934258577808 acquired on /root/.cache/huggingface/transformers/3c61d016573b14f7f008c02c4e51a366c67ab274726fe2910691e2a761acf43e.37395cee442ab11005bcd270f3c34464dc1704b715b5d7d52b1a461abe3b9e4e.lock
2021-06-01 19:42:32,444 INFO filelock Lock 139934258577808 released on /root/.cache/huggingface/transformers/3c61d016573b14f7f008c02c4e51a366c67ab274726fe2910691e2a761acf43e.37395cee442ab11005bcd270f3c34464dc1704b715b5d7d52b1a461abe3b9e4e.lock
2021-06-01 19:42:33,173 INFO filelock Lock 139934258580752 acquired on /root/.cache/huggingface/transformers/45c3f7a79a80e1cf0a489e5c62b43f173c15db47864303a55d623bb3c96f72a5.d789d64ebfe299b0e416afc4a169632f903f693095b4629a7ea271d5a0cf2c99.lock
2021-06-01 19:42:33,683 INFO filelock Lock 139934258580752 released on /root/.cache/huggingface/transformers/45c3f7a79a80e1cf0a489e5c62b43f173c15db47864303a55d623bb3c96f72a5.d789d64ebfe299b0e416afc4a169632f903f693095b4629a7ea271d5a0cf2c99.lock
2021-06-01 19:42:34,032 INFO filelock Lock 139934258102032 acquired on /root/.cache/huggingface/transformers/534479488c54aeaf9c3406f647aa2ec13648c06771ffe269edabebd4c412da1d.7f2721073f19841be16f41b0a70b600ca6b880c8f3df6f3535cbc704371bdfa4.lock
2021-06-01 19:42:34,565 INFO filelock Lock 139934258102032 released on /root/.cache/huggingface/transformers/534479488c54aeaf9c3406f647aa2ec13648c06771ffe269edabebd4c412da1d.7f2721073f19841be16f41b0a70b600ca6b880c8f3df6f3535cbc704371bdfa4.lock
2021-06-01 19:42:35,616 INFO filelock Lock 139934258119312 acquired on /root/.cache/huggingface/transformers/c1d7f0a763fb63861cc08553866f1fc3e5a6f4f07621be277452d26d71303b7e.20430bd8e10ef77a7d2977accefe796051e01bc2fc4aa146bc862997a1a15e79.lock
2021-06-01 19:42:35,975 INFO filelock Lock 139934258119312 released on /root/.cache/huggingface/transformers/c1d7f0a763fb63861cc08553866f1fc3e5a6f4f07621be277452d26d71303b7e.20430bd8e10ef77a7d2977accefe796051e01bc2fc4aa146bc862997a1a15e79.lock
2021-06-01 19:42:36,361 INFO filelock Lock 139934258578064 acquired on /root/.cache/huggingface/transformers/a8041bf617d7f94ea26d15e218abd04afc2004805632abc0ed2066aa16d50d04.faf6ea826ae9c5867d12b22257f9877e6b8367890837bd60f7c54a29633f7f2f.lock
2021-06-01 19:42:43,593 INFO filelock Lock 139934258578064 released on /root/.cache/huggingface/transformers/a8041bf617d7f94ea26d15e218abd04afc2004805632abc0ed2066aa16d50d04.faf6ea826ae9c5867d12b22257f9877e6b8367890837bd60f7c54a29633f7f2f.lock
2021-06-01 19:43:07,922 INFO __main__ Start training
2021-06-01 19:43:07,922 INFO torchdistill.models.util [teacher model]
2021-06-01 19:43:07,922 INFO torchdistill.models.util Using the original teacher model
2021-06-01 19:43:07,922 INFO torchdistill.models.util [student model]
2021-06-01 19:43:07,922 INFO torchdistill.models.util Using the original student model
2021-06-01 19:43:07,923 INFO torchdistill.core.distillation Loss = 1.0 * OrgLoss
2021-06-01 19:43:07,923 INFO torchdistill.core.distillation Freezing the whole teacher model
2021-06-01 19:43:14,731 INFO torchdistill.misc.log Epoch: [0] [ 0/3274] eta: 0:36:18 lr: 4.999490938709021e-05 sample/s: 6.268329780814906 loss: 0.6479 (0.6479) time: 0.6655 data: 0.0274 max mem: 3034
2021-06-01 19:48:36,357 INFO torchdistill.misc.log Epoch: [0] [ 500/3274] eta: 0:29:44 lr: 4.7449602932193036e-05 sample/s: 6.540947354960249 loss: 0.3303 (0.4136) time: 0.6375 data: 0.0044 max mem: 5082
2021-06-01 19:53:57,989 INFO torchdistill.misc.log Epoch: [0] [1000/3274] eta: 0:24:22 lr: 4.490429647729587e-05 sample/s: 6.062278385692094 loss: 0.2386 (0.3724) time: 0.6654 data: 0.0046 max mem: 5082
2021-06-01 19:59:20,881 INFO torchdistill.misc.log Epoch: [0] [1500/3274] eta: 0:19:02 lr: 4.23589900223987e-05 sample/s: 7.087861396219321 loss: 0.3091 (0.3486) time: 0.6669 data: 0.0044 max mem: 5082
2021-06-01 20:04:43,612 INFO torchdistill.misc.log Epoch: [0] [2000/3274] eta: 0:13:41 lr: 3.981368356750153e-05 sample/s: 6.062981631682493 loss: 0.2964 (0.3324) time: 0.6403 data: 0.0045 max mem: 5082
2021-06-01 20:10:04,214 INFO torchdistill.misc.log Epoch: [0] [2500/3274] eta: 0:08:18 lr: 3.726837711260436e-05 sample/s: 6.070009898869445 loss: 0.2783 (0.3190) time: 0.6220 data: 0.0045 max mem: 5082
2021-06-01 20:15:25,625 INFO torchdistill.misc.log Epoch: [0] [3000/3274] eta: 0:02:56 lr: 3.472307065770719e-05 sample/s: 6.05868802221069 loss: 0.2463 (0.3135) time: 0.6771 data: 0.0045 max mem: 5082
2021-06-01 20:18:20,347 INFO torchdistill.misc.log Epoch: [0] Total time: 0:35:06
2021-06-01 20:18:40,297 INFO /usr/local/lib/python3.7/dist-packages/datasets/metric.py Removing /root/.cache/huggingface/metrics/glue/qnli/default_experiment-1-0.arrow
2021-06-01 20:18:40,297 INFO __main__ Validation: accuracy = 0.90536335346879
2021-06-01 20:18:40,298 INFO __main__ Updating ckpt at ./resource/ckpt/glue/qnli/kd/qnli-bert-base-uncased_from_bert-large-uncased
2021-06-01 20:18:42,209 INFO torchdistill.misc.log Epoch: [1] [ 0/3274] eta: 0:32:03 lr: 3.332824272042354e-05 sample/s: 7.010533423814029 loss: 0.0507 (0.0507) time: 0.5874 data: 0.0169 max mem: 5082
2021-06-01 20:24:02,784 INFO torchdistill.misc.log Epoch: [1] [ 500/3274] eta: 0:29:38 lr: 3.078293626552637e-05 sample/s: 7.086708737782058 loss: 0.1098 (0.1673) time: 0.5952 data: 0.0044 max mem: 5082
2021-06-01 20:29:21,701 INFO torchdistill.misc.log Epoch: [1] [1000/3274] eta: 0:24:14 lr: 2.82376298106292e-05 sample/s: 6.536556439751819 loss: 0.1207 (0.1678) time: 0.6259 data: 0.0046 max mem: 5082
2021-06-01 20:34:42,169 INFO torchdistill.misc.log Epoch: [1] [1500/3274] eta: 0:18:55 lr: 2.569232335573203e-05 sample/s: 6.5383523579776845 loss: 0.0869 (0.1672) time: 0.6195 data: 0.0043 max mem: 5082
2021-06-01 20:40:06,680 INFO torchdistill.misc.log Epoch: [1] [2000/3274] eta: 0:13:38 lr: 2.314701690083486e-05 sample/s: 7.665929342032576 loss: 0.1216 (0.1649) time: 0.6682 data: 0.0044 max mem: 5082
2021-06-01 20:45:33,102 INFO torchdistill.misc.log Epoch: [1] [2500/3274] eta: 0:08:18 lr: 2.060171044593769e-05 sample/s: 7.08687637389096 loss: 0.1018 (0.1626) time: 0.6454 data: 0.0043 max mem: 5082
2021-06-01 20:50:54,681 INFO torchdistill.misc.log Epoch: [1] [3000/3274] eta: 0:02:56 lr: 1.8056403991040522e-05 sample/s: 7.090452503412688 loss: 0.0853 (0.1607) time: 0.6372 data: 0.0044 max mem: 5082
2021-06-01 20:53:50,080 INFO torchdistill.misc.log Epoch: [1] Total time: 0:35:08
2021-06-01 20:54:10,003 INFO /usr/local/lib/python3.7/dist-packages/datasets/metric.py Removing /root/.cache/huggingface/metrics/glue/qnli/default_experiment-1-0.arrow
2021-06-01 20:54:10,004 INFO __main__ Validation: accuracy = 0.9178107267069375
2021-06-01 20:54:10,004 INFO __main__ Updating ckpt at ./resource/ckpt/glue/qnli/kd/qnli-bert-base-uncased_from_bert-large-uncased
2021-06-01 20:54:11,975 INFO torchdistill.misc.log Epoch: [2] [ 0/3274] eta: 0:32:07 lr: 1.6661576053756872e-05 sample/s: 6.9625433519294795 loss: 0.0285 (0.0285) time: 0.5887 data: 0.0142 max mem: 5082
2021-06-01 20:59:31,890 INFO torchdistill.misc.log Epoch: [2] [ 500/3274] eta: 0:29:34 lr: 1.4116269598859703e-05 sample/s: 6.541477823907571 loss: 0.0000 (0.0754) time: 0.6146 data: 0.0045 max mem: 5082
2021-06-01 21:04:51,984 INFO torchdistill.misc.log Epoch: [2] [1000/3274] eta: 0:24:15 lr: 1.1570963143962533e-05 sample/s: 7.666114992529096 loss: 0.0842 (0.1092) time: 0.6460 data: 0.0046 max mem: 5082
2021-06-01 21:10:11,480 INFO torchdistill.misc.log Epoch: [2] [1500/3274] eta: 0:18:54 lr: 9.025656689065364e-06 sample/s: 5.0033478488296845 loss: 0.0114 (0.1235) time: 0.6660 data: 0.0044 max mem: 5082
2021-06-01 21:15:33,085 INFO torchdistill.misc.log Epoch: [2] [2000/3274] eta: 0:13:36 lr: 6.480350234168193e-06 sample/s: 5.000420546137726 loss: 0.0001 (0.1282) time: 0.6606 data: 0.0044 max mem: 5082
2021-06-01 21:20:54,238 INFO torchdistill.misc.log Epoch: [2] [2500/3274] eta: 0:08:16 lr: 3.935043779271024e-06 sample/s: 7.088882635439985 loss: 0.0000 (0.1301) time: 0.5969 data: 0.0042 max mem: 5082
2021-06-01 21:26:15,307 INFO torchdistill.misc.log Epoch: [2] [3000/3274] eta: 0:02:55 lr: 1.3897373243738546e-06 sample/s: 6.539550185149094 loss: 0.0000 (0.1303) time: 0.6694 data: 0.0045 max mem: 5082
2021-06-01 21:29:10,621 INFO torchdistill.misc.log Epoch: [2] Total time: 0:34:59
2021-06-01 21:29:30,567 INFO /usr/local/lib/python3.7/dist-packages/datasets/metric.py Removing /root/.cache/huggingface/metrics/glue/qnli/default_experiment-1-0.arrow
2021-06-01 21:29:30,567 INFO __main__ Validation: accuracy = 0.9101226432363171
2021-06-01 21:29:30,604 INFO __main__ [Teacher: bert-large-uncased]
2021-06-01 21:30:27,885 INFO /usr/local/lib/python3.7/dist-packages/datasets/metric.py Removing /root/.cache/huggingface/metrics/glue/qnli/default_experiment-1-0.arrow
2021-06-01 21:30:27,886 INFO __main__ Test: accuracy = 0.9223869668680212
2021-06-01 21:30:31,466 INFO __main__ [Student: bert-base-uncased]
2021-06-01 21:30:51,406 INFO /usr/local/lib/python3.7/dist-packages/datasets/metric.py Removing /root/.cache/huggingface/metrics/glue/qnli/default_experiment-1-0.arrow
2021-06-01 21:30:51,406 INFO __main__ Test: accuracy = 0.9178107267069375
2021-06-01 21:30:51,406 INFO __main__ Start prediction for private dataset(s)
2021-06-01 21:30:51,407 INFO __main__ qnli/test: 5463 samples