sanchit-gandhi HF staff commited on
Commit
4a6a658
1 Parent(s): 79493fa

Training in progress, step 500

Browse files
.gitattributes CHANGED
@@ -28,3 +28,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
28
  wandb/run-20220228_093705-yn2gmwrw/run-yn2gmwrw.wandb filter=lfs diff=lfs merge=lfs -text
29
  wandb/run-20220228_122826-3vepvy3m/run-3vepvy3m.wandb filter=lfs diff=lfs merge=lfs -text
30
  wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb filter=lfs diff=lfs merge=lfs -text
 
 
28
  wandb/run-20220228_093705-yn2gmwrw/run-yn2gmwrw.wandb filter=lfs diff=lfs merge=lfs -text
29
  wandb/run-20220228_122826-3vepvy3m/run-3vepvy3m.wandb filter=lfs diff=lfs merge=lfs -text
30
  wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb filter=lfs diff=lfs merge=lfs -text
31
+ wandb/run-20220228_163348-20lja62k/run-20lja62k.wandb filter=lfs diff=lfs merge=lfs -text
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:692c5cce6d3639f3cd8462fa2ea669b75c2c3e0a0a1fa5c036ef54ae18f470b6
3
  size 3210531882
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4fb569df25b38322a3b551ceca0b840fe40dbb8b8287a64c5d584c47e743372b
3
  size 3210531882
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:2d6def6b2adb48124aa28e6a00999430af85df3f733e9e528f2093f433a40ff6
3
  size 3119
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:afef645ecdffc5ee7693a084c9c81f9caea1a19faff024b129d7f6503a77bf58
3
  size 3119
wandb/debug-internal.log CHANGED
@@ -1 +1 @@
1
- run-20220228_144113-18osrjm3/logs/debug-internal.log
 
1
+ run-20220228_163348-20lja62k/logs/debug-internal.log
wandb/debug.log CHANGED
@@ -1 +1 @@
1
- run-20220228_144113-18osrjm3/logs/debug.log
 
1
+ run-20220228_163348-20lja62k/logs/debug.log
wandb/latest-run CHANGED
@@ -1 +1 @@
1
- run-20220228_144113-18osrjm3
 
1
+ run-20220228_163348-20lja62k
wandb/run-20220228_144113-18osrjm3/files/config.yaml CHANGED
@@ -10673,7 +10673,14 @@ _wandb:
10673
  - 1
10674
  - 5
10675
  - 11
 
 
 
 
 
10676
  3:
 
 
10677
  - 13
10678
  4: 3.9.5
10679
  5: 0.12.10
 
10673
  - 1
10674
  - 5
10675
  - 11
10676
+ 2:
10677
+ - 1
10678
+ - 5
10679
+ - 11
10680
+ - 12
10681
  3:
10682
+ - 1
10683
+ - 7
10684
  - 13
10685
  4: 3.9.5
10686
  5: 0.12.10
wandb/run-20220228_144113-18osrjm3/files/output.log CHANGED
@@ -3497,3 +3497,10 @@ Upload file wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb: 100%|██
3497
  eval_steps_per_second = 0.292
3498
  100%|█████████████████████████████████████████████████████████████████████████████████| 331/331 [18:50<00:00, 3.41s/it]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3499
  [INFO|modeling_utils.py:1081] 2022-02-28 16:31:50,167 >> Model weights saved in ./pytorch_model.bin:50<00:00, 3.41s/it]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
 
 
 
 
 
 
 
 
3497
  eval_steps_per_second = 0.292
3498
  100%|█████████████████████████████████████████████████████████████████████████████████| 331/331 [18:50<00:00, 3.41s/it]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3499
  [INFO|modeling_utils.py:1081] 2022-02-28 16:31:50,167 >> Model weights saved in ./pytorch_model.bin:50<00:00, 3.41s/it]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3500
+ Upload file wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb: 26%|██▊ | 14.0M/53.8M [00:01<00:02, 14.7MB/s]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3501
+ Upload file wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb: 100%|███████████| 53.8M/53.8M [00:03<00:00, 18.7MB/s]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3502
+ Upload file wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb: 100%|███████████| 53.8M/53.8M [00:03<00:00, 18.7MB/s]To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3503
+ 02/28/2022 16:32:20 - WARNING - huggingface_hub.repository - To https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-search
3504
+ File "/home/sanchit_huggingface_co/gcp/lib/python3.9/site-packages/huggingface_hub/hf_api.py", line 870, in model_infoTo https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3505
+ File "/home/sanchit_huggingface_co/gcp/lib/python3.9/site-packages/huggingface_hub/hf_api.py", line 870, in model_infoTo https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
3506
+ File "/home/sanchit_huggingface_co/gcp/lib/python3.9/site-packages/huggingface_hub/hf_api.py", line 870, in model_infoTo https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-searchimate the number of tokens of the input, floating-point operations will not be computed
wandb/run-20220228_144113-18osrjm3/files/wandb-summary.json CHANGED
The diff for this file is too large to render. See raw diff
 
wandb/run-20220228_144113-18osrjm3/logs/debug-internal.log CHANGED
@@ -8732,3 +8732,117 @@
8732
  2022-02-28 16:31:55,743 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: stop_status
8733
  2022-02-28 16:32:10,787 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: stop_status
8734
  2022-02-28 16:32:10,788 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: stop_status
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8732
  2022-02-28 16:31:55,743 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: stop_status
8733
  2022-02-28 16:32:10,787 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: stop_status
8734
  2022-02-28 16:32:10,788 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: stop_status
8735
+ 2022-02-28 16:32:17,396 DEBUG SenderThread:224575 [sender.py:send():235] send: stats
8736
+ 2022-02-28 16:32:19,709 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log
8737
+ 2022-02-28 16:32:21,710 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log
8738
+ 2022-02-28 16:32:22,710 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log
8739
+ 2022-02-28 16:32:25,711 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log
8740
+ 2022-02-28 16:32:26,336 DEBUG SenderThread:224575 [sender.py:send():235] send: telemetry
8741
+ 2022-02-28 16:32:26,336 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8742
+ 2022-02-28 16:32:26,336 DEBUG SenderThread:224575 [sender.py:send():235] send: exit
8743
+ 2022-02-28 16:32:26,336 INFO SenderThread:224575 [sender.py:send_exit():371] handling exit code: 1
8744
+ 2022-02-28 16:32:26,337 INFO SenderThread:224575 [sender.py:send_exit():373] handling runtime: 6671
8745
+ 2022-02-28 16:32:26,391 INFO SenderThread:224575 [sender.py:_save_file():944] saving file wandb-summary.json with policy end
8746
+ 2022-02-28 16:32:26,392 INFO SenderThread:224575 [sender.py:send_exit():379] send defer
8747
+ 2022-02-28 16:32:26,392 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8748
+ 2022-02-28 16:32:26,392 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8749
+ 2022-02-28 16:32:26,392 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 0
8750
+ 2022-02-28 16:32:26,393 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8751
+ 2022-02-28 16:32:26,393 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 0
8752
+ 2022-02-28 16:32:26,393 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 1
8753
+ 2022-02-28 16:32:26,393 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8754
+ 2022-02-28 16:32:26,393 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 1
8755
+ 2022-02-28 16:32:26,510 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8756
+ 2022-02-28 16:32:26,510 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 1
8757
+ 2022-02-28 16:32:26,510 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 2
8758
+ 2022-02-28 16:32:26,511 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8759
+ 2022-02-28 16:32:26,511 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8760
+ 2022-02-28 16:32:26,512 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8761
+ 2022-02-28 16:32:26,512 DEBUG SenderThread:224575 [sender.py:send():235] send: stats
8762
+ 2022-02-28 16:32:26,512 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 2
8763
+ 2022-02-28 16:32:26,513 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8764
+ 2022-02-28 16:32:26,513 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 2
8765
+ 2022-02-28 16:32:26,513 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 3
8766
+ 2022-02-28 16:32:26,513 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8767
+ 2022-02-28 16:32:26,513 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 3
8768
+ 2022-02-28 16:32:26,572 DEBUG SenderThread:224575 [sender.py:send():235] send: summary
8769
+ 2022-02-28 16:32:26,661 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8770
+ 2022-02-28 16:32:26,663 INFO SenderThread:224575 [sender.py:_save_file():944] saving file wandb-summary.json with policy end
8771
+ 2022-02-28 16:32:26,664 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8772
+ 2022-02-28 16:32:26,664 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 3
8773
+ 2022-02-28 16:32:26,664 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 4
8774
+ 2022-02-28 16:32:26,664 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8775
+ 2022-02-28 16:32:26,665 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8776
+ 2022-02-28 16:32:26,665 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 4
8777
+ 2022-02-28 16:32:26,665 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8778
+ 2022-02-28 16:32:26,665 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 4
8779
+ 2022-02-28 16:32:26,711 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/wandb-summary.json
8780
+ 2022-02-28 16:32:26,711 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log
8781
+ 2022-02-28 16:32:26,766 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8782
+ 2022-02-28 16:32:27,491 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 5
8783
+ 2022-02-28 16:32:27,492 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8784
+ 2022-02-28 16:32:27,492 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8785
+ 2022-02-28 16:32:27,493 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 5
8786
+ 2022-02-28 16:32:27,493 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8787
+ 2022-02-28 16:32:27,493 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 5
8788
+ 2022-02-28 16:32:27,493 INFO SenderThread:224575 [dir_watcher.py:finish():283] shutting down directory watcher
8789
+ 2022-02-28 16:32:27,593 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8790
+ 2022-02-28 16:32:27,712 INFO Thread-8 :224575 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/config.yaml
8791
+ 2022-02-28 16:32:27,712 INFO SenderThread:224575 [dir_watcher.py:finish():313] scan: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files
8792
+ 2022-02-28 16:32:27,712 INFO SenderThread:224575 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/wandb-metadata.json wandb-metadata.json
8793
+ 2022-02-28 16:32:27,712 INFO SenderThread:224575 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log output.log
8794
+ 2022-02-28 16:32:27,712 INFO SenderThread:224575 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/wandb-summary.json wandb-summary.json
8795
+ 2022-02-28 16:32:27,713 INFO SenderThread:224575 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/requirements.txt requirements.txt
8796
+ 2022-02-28 16:32:27,715 INFO SenderThread:224575 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/config.yaml config.yaml
8797
+ 2022-02-28 16:32:27,716 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 6
8798
+ 2022-02-28 16:32:27,716 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8799
+ 2022-02-28 16:32:27,716 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8800
+ 2022-02-28 16:32:27,716 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 6
8801
+ 2022-02-28 16:32:27,717 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8802
+ 2022-02-28 16:32:27,717 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 6
8803
+ 2022-02-28 16:32:27,717 INFO SenderThread:224575 [file_pusher.py:finish():177] shutting down file pusher
8804
+ 2022-02-28 16:32:27,818 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8805
+ 2022-02-28 16:32:27,818 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8806
+ 2022-02-28 16:32:27,920 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8807
+ 2022-02-28 16:32:27,920 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8808
+ 2022-02-28 16:32:28,022 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8809
+ 2022-02-28 16:32:28,022 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8810
+ 2022-02-28 16:32:28,032 INFO Thread-14 :224575 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/requirements.txt
8811
+ 2022-02-28 16:32:28,035 INFO Thread-12 :224575 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/output.log
8812
+ 2022-02-28 16:32:28,047 INFO Thread-15 :224575 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/config.yaml
8813
+ 2022-02-28 16:32:28,096 INFO Thread-13 :224575 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/files/wandb-summary.json
8814
+ 2022-02-28 16:32:28,123 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8815
+ 2022-02-28 16:32:28,124 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8816
+ 2022-02-28 16:32:28,225 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8817
+ 2022-02-28 16:32:28,225 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8818
+ 2022-02-28 16:32:28,296 INFO Thread-7 :224575 [sender.py:transition_state():392] send defer: 7
8819
+ 2022-02-28 16:32:28,297 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8820
+ 2022-02-28 16:32:28,297 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 7
8821
+ 2022-02-28 16:32:28,297 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8822
+ 2022-02-28 16:32:28,297 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 7
8823
+ 2022-02-28 16:32:28,327 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8824
+ 2022-02-28 16:32:29,749 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 8
8825
+ 2022-02-28 16:32:29,749 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8826
+ 2022-02-28 16:32:29,750 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8827
+ 2022-02-28 16:32:29,750 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 8
8828
+ 2022-02-28 16:32:29,750 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8829
+ 2022-02-28 16:32:29,750 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 8
8830
+ 2022-02-28 16:32:29,751 INFO SenderThread:224575 [sender.py:transition_state():392] send defer: 9
8831
+ 2022-02-28 16:32:29,751 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: defer
8832
+ 2022-02-28 16:32:29,752 DEBUG SenderThread:224575 [sender.py:send():235] send: final
8833
+ 2022-02-28 16:32:29,752 INFO HandlerThread:224575 [handler.py:handle_request_defer():154] handle defer: 9
8834
+ 2022-02-28 16:32:29,753 DEBUG SenderThread:224575 [sender.py:send():235] send: footer
8835
+ 2022-02-28 16:32:29,753 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: defer
8836
+ 2022-02-28 16:32:29,753 INFO SenderThread:224575 [sender.py:send_request_defer():388] handle sender defer: 9
8837
+ 2022-02-28 16:32:29,851 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: poll_exit
8838
+ 2022-02-28 16:32:29,851 DEBUG SenderThread:224575 [sender.py:send_request():249] send_request: poll_exit
8839
+ 2022-02-28 16:32:29,852 INFO SenderThread:224575 [file_pusher.py:join():182] waiting for file pusher
8840
+ 2022-02-28 16:32:29,922 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: get_summary
8841
+ 2022-02-28 16:32:30,033 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: sampled_history
8842
+ 2022-02-28 16:32:30,036 DEBUG HandlerThread:224575 [handler.py:handle_request():131] handle_request: shutdown
8843
+ 2022-02-28 16:32:30,037 INFO HandlerThread:224575 [handler.py:finish():739] shutting down handler
8844
+ 2022-02-28 16:32:30,752 INFO WriterThread:224575 [datastore.py:close():281] close: /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb
8845
+ 2022-02-28 16:32:30,921 INFO SenderThread:224575 [sender.py:finish():1075] shutting down sender
8846
+ 2022-02-28 16:32:30,921 INFO SenderThread:224575 [file_pusher.py:finish():177] shutting down file pusher
8847
+ 2022-02-28 16:32:30,922 INFO SenderThread:224575 [file_pusher.py:join():182] waiting for file pusher
8848
+ 2022-02-28 16:32:30,931 INFO MainThread:224575 [internal.py:handle_exit():79] Internal process exited
wandb/run-20220228_144113-18osrjm3/logs/debug.log CHANGED
@@ -24,3 +24,109 @@ config: {}
24
  2022-02-28 14:41:14,472 INFO MainThread:224433 [wandb_init.py:init():651] run started, returning control to user process
25
  2022-02-28 14:41:14,474 INFO MainThread:224433 [wandb_run.py:_config_callback():966] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'torch.float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': False, 'is_encoder_decoder': True, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 50, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['SpeechEncoderDecoderModel'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 50256, 'eos_token_id': 50256, 'sep_token_id': None, 'decoder_start_token_id': 50256, 'task_specific_params': None, 'problem_type': None, '_name_or_path': './', 'transformers_version': None, 'decoder': {'vocab_size': 50257, 'n_positions': 1024, 'n_embd': 1024, 'n_layer': 24, 'n_head': 16, 'n_inner': None, 'activation_function': 'gelu_new', 'resid_pdrop': 0.0, 'embd_pdrop': 0.0, 'attn_pdrop': 0.0, 'layer_norm_epsilon': 1e-05, 'initializer_range': 0.02, 'summary_type': 'cls_index', 'summary_use_proj': True, 'summary_activation': None, 'summary_first_dropout': 0.0, 'summary_proj_to_labels': True, 'scale_attn_weights': True, 'use_cache': False, 'scale_attn_by_inverse_layer_idx': False, 'reorder_and_upcast_attn': False, 'bos_token_id': 50256, 'eos_token_id': 50256, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': True, 'cross_attention_hidden_size': None, 'add_cross_attention': True, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['GPT2LMHeadModel'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'pad_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': {'text-generation': {'do_sample': True, 'max_length': 50}}, 'problem_type': None, '_name_or_path': 'gpt2-medium', 'transformers_version': '4.17.0.dev0', 'n_ctx': 1024, 'n_special': 0, 'predict_special_tokens': True, 'model_type': 'gpt2'}, 'encoder': {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 0, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-large-lv60', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'gradient_checkpointing': False, 'hidden_dropout_prob': 0.0, 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.0, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 32, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': False, 'mask_time_prob': 0.0, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.0, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'sum', 'ctc_zero_infinity': False, 'add_adapter': True, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'model_type': 'wav2vec2'}, 'model_type': 'speech-encoder-decoder', 'processor_class': 'Wav2Vec2Processor', 'use_cache': False, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 8, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 8, 'eval_accumulation_steps': 'None', 'learning_rate': 1e-05, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 1.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Feb28_14-40-31_sanchit--v100', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 1, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 1, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': './', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': False, 'metric_for_best_model': 'None', 'greater_is_better': 'None', 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'sortish_sampler': False, 'predict_with_generate': True, 'generation_max_length': 40, 'generation_num_beams': 1, 'train_batch_size': 8, 'eval_batch_size': 8}
26
  2022-02-28 14:41:14,477 INFO MainThread:224433 [wandb_watch.py:watch():43] Watching
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
24
  2022-02-28 14:41:14,472 INFO MainThread:224433 [wandb_init.py:init():651] run started, returning control to user process
25
  2022-02-28 14:41:14,474 INFO MainThread:224433 [wandb_run.py:_config_callback():966] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'torch.float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': False, 'is_encoder_decoder': True, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 50, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['SpeechEncoderDecoderModel'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 50256, 'eos_token_id': 50256, 'sep_token_id': None, 'decoder_start_token_id': 50256, 'task_specific_params': None, 'problem_type': None, '_name_or_path': './', 'transformers_version': None, 'decoder': {'vocab_size': 50257, 'n_positions': 1024, 'n_embd': 1024, 'n_layer': 24, 'n_head': 16, 'n_inner': None, 'activation_function': 'gelu_new', 'resid_pdrop': 0.0, 'embd_pdrop': 0.0, 'attn_pdrop': 0.0, 'layer_norm_epsilon': 1e-05, 'initializer_range': 0.02, 'summary_type': 'cls_index', 'summary_use_proj': True, 'summary_activation': None, 'summary_first_dropout': 0.0, 'summary_proj_to_labels': True, 'scale_attn_weights': True, 'use_cache': False, 'scale_attn_by_inverse_layer_idx': False, 'reorder_and_upcast_attn': False, 'bos_token_id': 50256, 'eos_token_id': 50256, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': True, 'cross_attention_hidden_size': None, 'add_cross_attention': True, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['GPT2LMHeadModel'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'pad_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': {'text-generation': {'do_sample': True, 'max_length': 50}}, 'problem_type': None, '_name_or_path': 'gpt2-medium', 'transformers_version': '4.17.0.dev0', 'n_ctx': 1024, 'n_special': 0, 'predict_special_tokens': True, 'model_type': 'gpt2'}, 'encoder': {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 0, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-large-lv60', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'gradient_checkpointing': False, 'hidden_dropout_prob': 0.0, 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.0, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 32, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': False, 'mask_time_prob': 0.0, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.0, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'sum', 'ctc_zero_infinity': False, 'add_adapter': True, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'model_type': 'wav2vec2'}, 'model_type': 'speech-encoder-decoder', 'processor_class': 'Wav2Vec2Processor', 'use_cache': False, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 8, 'per_device_eval_batch_size': 8, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 8, 'eval_accumulation_steps': 'None', 'learning_rate': 1e-05, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 1.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Feb28_14-40-31_sanchit--v100', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 1, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 1, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': './', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': False, 'metric_for_best_model': 'None', 'greater_is_better': 'None', 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'sortish_sampler': False, 'predict_with_generate': True, 'generation_max_length': 40, 'generation_num_beams': 1, 'train_batch_size': 8, 'eval_batch_size': 8}
26
  2022-02-28 14:41:14,477 INFO MainThread:224433 [wandb_watch.py:watch():43] Watching
27
+ 2022-02-28 16:32:24,188 INFO MainThread:224433 [wandb_run.py:_atexit_cleanup():1797] got exitcode: 1
28
+ 2022-02-28 16:32:24,191 INFO MainThread:224433 [wandb_run.py:_restore():1769] restore
29
+ 2022-02-28 16:32:26,392 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
30
+ wandb_count: 1
31
+ }
32
+ pusher_stats {
33
+ uploaded_bytes: 2093
34
+ total_bytes: 2093
35
+ }
36
+
37
+ 2022-02-28 16:32:26,512 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
38
+ wandb_count: 1
39
+ }
40
+ pusher_stats {
41
+ uploaded_bytes: 2093
42
+ total_bytes: 2093
43
+ }
44
+
45
+ 2022-02-28 16:32:26,665 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
46
+ wandb_count: 1
47
+ }
48
+ pusher_stats {
49
+ uploaded_bytes: 2093
50
+ total_bytes: 2093
51
+ }
52
+
53
+ 2022-02-28 16:32:27,493 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
54
+ wandb_count: 1
55
+ }
56
+ pusher_stats {
57
+ uploaded_bytes: 2093
58
+ total_bytes: 2093
59
+ }
60
+
61
+ 2022-02-28 16:32:27,717 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
62
+ wandb_count: 2
63
+ }
64
+ pusher_stats {
65
+ uploaded_bytes: 2093
66
+ total_bytes: 972408
67
+ }
68
+
69
+ 2022-02-28 16:32:27,819 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
70
+ wandb_count: 5
71
+ }
72
+ pusher_stats {
73
+ uploaded_bytes: 2093
74
+ total_bytes: 2832309
75
+ }
76
+
77
+ 2022-02-28 16:32:27,921 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
78
+ wandb_count: 5
79
+ }
80
+ pusher_stats {
81
+ uploaded_bytes: 2832309
82
+ total_bytes: 2832309
83
+ }
84
+
85
+ 2022-02-28 16:32:28,022 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
86
+ wandb_count: 5
87
+ }
88
+ pusher_stats {
89
+ uploaded_bytes: 2832309
90
+ total_bytes: 2832309
91
+ }
92
+
93
+ 2022-02-28 16:32:28,124 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
94
+ wandb_count: 5
95
+ }
96
+ pusher_stats {
97
+ uploaded_bytes: 2832309
98
+ total_bytes: 2832309
99
+ }
100
+
101
+ 2022-02-28 16:32:28,226 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
102
+ wandb_count: 5
103
+ }
104
+ pusher_stats {
105
+ uploaded_bytes: 2832309
106
+ total_bytes: 2832309
107
+ }
108
+
109
+ 2022-02-28 16:32:29,750 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
110
+ wandb_count: 5
111
+ }
112
+ pusher_stats {
113
+ uploaded_bytes: 2832309
114
+ total_bytes: 2832309
115
+ }
116
+
117
+ 2022-02-28 16:32:29,922 INFO MainThread:224433 [wandb_run.py:_wait_for_finish():1929] got exit ret: done: true
118
+ exit_result {
119
+ }
120
+ file_counts {
121
+ wandb_count: 5
122
+ }
123
+ pusher_stats {
124
+ uploaded_bytes: 2832309
125
+ total_bytes: 2832309
126
+ }
127
+ local_info {
128
+ }
129
+
130
+ 2022-02-28 16:32:31,095 INFO MainThread:224433 [wandb_run.py:_append_history():2144] rendering history
131
+ 2022-02-28 16:32:31,096 INFO MainThread:224433 [wandb_run.py:_append_summary():2102] rendering summary
132
+ 2022-02-28 16:32:31,097 INFO MainThread:224433 [wandb_run.py:_append_files():2194] logging synced files
wandb/run-20220228_144113-18osrjm3/run-18osrjm3.wandb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:85d90845ebb54a0cb6b279d3bcdf721a69f247791dd753a9db94abf736be32a1
3
- size 56393843
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:64bf5a8eb58723501242a4453ad3cf1a1c63f51ae2299de26af7a72ef4e56bab
3
+ size 57959276
wandb/run-20220228_163348-20lja62k/files/config.yaml ADDED
The diff for this file is too large to render. See raw diff
 
wandb/run-20220228_163348-20lja62k/files/output.log ADDED
The diff for this file is too large to render. See raw diff
 
wandb/run-20220228_163348-20lja62k/files/requirements.txt ADDED
@@ -0,0 +1,184 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ absl-py==1.0.0
2
+ aiohttp==3.8.1
3
+ aiosignal==1.2.0
4
+ anyio==3.5.0
5
+ appdirs==1.4.4
6
+ argon2-cffi-bindings==21.2.0
7
+ argon2-cffi==21.3.0
8
+ asttokens==2.0.5
9
+ async-timeout==4.0.2
10
+ attrs==21.4.0
11
+ audioread==2.1.9
12
+ babel==2.9.1
13
+ backcall==0.2.0
14
+ bitsandbytes-cuda113==0.26.0
15
+ black==22.1.0
16
+ bleach==4.1.0
17
+ cachetools==5.0.0
18
+ certifi==2021.10.8
19
+ cffi==1.15.0
20
+ charset-normalizer==2.0.11
21
+ chex==0.1.0
22
+ click==8.0.3
23
+ clldutils==3.10.1
24
+ colorlog==6.6.0
25
+ csvw==1.11.0
26
+ cycler==0.11.0
27
+ datasets==1.18.3
28
+ debugpy==1.5.1
29
+ decorator==5.1.1
30
+ defusedxml==0.7.1
31
+ dill==0.3.4
32
+ dlinfo==1.2.1
33
+ dm-tree==0.1.6
34
+ docker-pycreds==0.4.0
35
+ entrypoints==0.4
36
+ executing==0.8.2
37
+ filelock==3.4.2
38
+ flatbuffers==2.0
39
+ flax==0.4.0
40
+ fonttools==4.29.1
41
+ frozenlist==1.3.0
42
+ fsspec==2022.1.0
43
+ gitdb==4.0.9
44
+ gitpython==3.1.27
45
+ google-auth-oauthlib==0.4.6
46
+ google-auth==2.6.0
47
+ grpcio==1.43.0
48
+ huggingface-hub==0.4.0
49
+ hypothesis==6.36.1
50
+ idna==3.3
51
+ importlib-metadata==4.10.1
52
+ ipykernel==6.8.0
53
+ ipython-genutils==0.2.0
54
+ ipython==8.0.1
55
+ ipywidgets==7.6.5
56
+ isodate==0.6.1
57
+ jax==0.2.28
58
+ jaxlib==0.1.76+cuda11.cudnn82
59
+ jedi==0.18.1
60
+ jinja2==3.0.3
61
+ jiwer==2.3.0
62
+ joblib==1.1.0
63
+ json5==0.9.6
64
+ jsonschema==4.4.0
65
+ jupyter-client==7.1.2
66
+ jupyter-console==6.4.0
67
+ jupyter-core==4.9.1
68
+ jupyter-server==1.13.5
69
+ jupyter==1.0.0
70
+ jupyterlab-pygments==0.1.2
71
+ jupyterlab-server==2.10.3
72
+ jupyterlab-widgets==1.0.2
73
+ jupyterlab==3.2.9
74
+ kiwisolver==1.3.2
75
+ librosa==0.8.1
76
+ llvmlite==0.38.0
77
+ markdown==3.3.6
78
+ markupsafe==2.0.1
79
+ matplotlib-inline==0.1.3
80
+ matplotlib==3.5.1
81
+ mistune==0.8.4
82
+ msgpack==1.0.3
83
+ multidict==6.0.2
84
+ multiprocess==0.70.12.2
85
+ mypy-extensions==0.4.3
86
+ nbclassic==0.3.5
87
+ nbclient==0.5.10
88
+ nbconvert==6.4.1
89
+ nbformat==5.1.3
90
+ nest-asyncio==1.5.4
91
+ notebook==6.4.8
92
+ numba==0.55.1
93
+ numpy==1.21.5
94
+ oauthlib==3.2.0
95
+ opt-einsum==3.3.0
96
+ optax==0.1.0
97
+ packaging==21.3
98
+ pandas==1.4.0
99
+ pandocfilters==1.5.0
100
+ parso==0.8.3
101
+ pathspec==0.9.0
102
+ pathtools==0.1.2
103
+ pexpect==4.8.0
104
+ phonemizer==3.0.1
105
+ pickleshare==0.7.5
106
+ pillow==9.0.0
107
+ pip==22.0.2
108
+ pkg-resources==0.0.0
109
+ platformdirs==2.4.1
110
+ pooch==1.6.0
111
+ prometheus-client==0.13.1
112
+ promise==2.3
113
+ prompt-toolkit==3.0.26
114
+ protobuf==3.19.4
115
+ psutil==5.9.0
116
+ ptyprocess==0.7.0
117
+ pure-eval==0.2.2
118
+ pyarrow==6.0.1
119
+ pyasn1-modules==0.2.8
120
+ pyasn1==0.4.8
121
+ pycparser==2.21
122
+ pyctcdecode==0.3.0
123
+ pygments==2.11.2
124
+ pygtrie==2.4.2
125
+ pyparsing==3.0.7
126
+ pyrsistent==0.18.1
127
+ python-dateutil==2.8.2
128
+ python-levenshtein==0.12.2
129
+ pytz==2021.3
130
+ pyyaml==6.0
131
+ pyzmq==22.3.0
132
+ qtconsole==5.2.2
133
+ qtpy==2.0.1
134
+ regex==2022.1.18
135
+ requests-oauthlib==1.3.1
136
+ requests==2.27.1
137
+ resampy==0.2.2
138
+ rfc3986==2.0.0
139
+ rsa==4.8
140
+ sacremoses==0.0.47
141
+ scikit-learn==1.0.2
142
+ scipy==1.7.3
143
+ segments==2.2.0
144
+ send2trash==1.8.0
145
+ sentry-sdk==1.5.6
146
+ setuptools==44.1.1
147
+ shortuuid==1.0.8
148
+ six==1.16.0
149
+ smmap==5.0.0
150
+ sniffio==1.2.0
151
+ sortedcontainers==2.4.0
152
+ soundfile==0.10.3.post1
153
+ stack-data==0.1.4
154
+ tabulate==0.8.9
155
+ tensorboard-data-server==0.6.1
156
+ tensorboard-plugin-wit==1.8.1
157
+ tensorboard==2.8.0
158
+ termcolor==1.1.0
159
+ terminado==0.13.1
160
+ testpath==0.5.0
161
+ threadpoolctl==3.1.0
162
+ tokenizers==0.11.4
163
+ tomli==2.0.0
164
+ toolz==0.11.2
165
+ torch==1.10.2+cu113
166
+ torchaudio==0.10.2+cu113
167
+ tornado==6.1
168
+ tqdm==4.62.3
169
+ traitlets==5.1.1
170
+ transformers==4.17.0.dev0
171
+ typing-extensions==3.10.0.2
172
+ uritemplate==4.1.1
173
+ urllib3==1.26.8
174
+ wandb==0.12.10
175
+ wcwidth==0.2.5
176
+ webencodings==0.5.1
177
+ websocket-client==1.2.3
178
+ werkzeug==2.0.2
179
+ wheel==0.37.1
180
+ widgetsnbextension==3.5.2
181
+ xxhash==2.0.2
182
+ yarl==1.7.2
183
+ yaspin==2.1.0
184
+ zipp==3.7.0
wandb/run-20220228_163348-20lja62k/files/wandb-metadata.json ADDED
@@ -0,0 +1,60 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-5.11.0-1028-gcp-x86_64-with-glibc2.33",
3
+ "python": "3.9.5",
4
+ "heartbeatAt": "2022-02-28T16:33:50.087376",
5
+ "startedAt": "2022-02-28T16:33:48.958666",
6
+ "docker": null,
7
+ "gpu": "Tesla V100-SXM2-16GB",
8
+ "gpu_count": 2,
9
+ "cpu_count": 16,
10
+ "cuda": null,
11
+ "args": [
12
+ "--dataset_name=librispeech_asr",
13
+ "--model_name_or_path=./",
14
+ "--tokenizer_name=./",
15
+ "--dataset_config_name=clean",
16
+ "--train_split_name=train.100",
17
+ "--eval_split_name=validation",
18
+ "--output_dir=./",
19
+ "--preprocessing_num_workers=1",
20
+ "--length_column_name=input_length",
21
+ "--overwrite_output_dir",
22
+ "--num_train_epochs=1",
23
+ "--per_device_train_batch_size=12",
24
+ "--per_device_eval_batch_size=12",
25
+ "--gradient_accumulation_steps=2",
26
+ "--generation_max_length=40",
27
+ "--generation_num_beams=1",
28
+ "--learning_rate=1e-5",
29
+ "--warmup_steps=500",
30
+ "--evaluation_strategy=steps",
31
+ "--text_column_name=text",
32
+ "--save_steps=500",
33
+ "--eval_steps=500",
34
+ "--logging_steps=1",
35
+ "--save_total_limit=1",
36
+ "--freeze_feature_encoder",
37
+ "--gradient_checkpointing",
38
+ "--fp16",
39
+ "--group_by_length",
40
+ "--predict_with_generate",
41
+ "--do_lower_case",
42
+ "--do_train",
43
+ "--do_eval",
44
+ "--report_to=wandb",
45
+ "--push_to_hub",
46
+ "--use_auth_token"
47
+ ],
48
+ "state": "running",
49
+ "program": "/home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/run_speech_recognition_seq2seq.py",
50
+ "codePath": "run_speech_recognition_seq2seq.py",
51
+ "git": {
52
+ "remote": "https://huggingface.co/sanchit-gandhi/wav2vec2-gpt2-wandb-grid-search",
53
+ "commit": "79493face58ddfbd26180e47b578274dbeaeed22"
54
+ },
55
+ "email": "sanchit@huggingface.co",
56
+ "root": "/home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search",
57
+ "host": "sanchit--v100",
58
+ "username": "sanchit_huggingface_co",
59
+ "executable": "/home/sanchit_huggingface_co/gcp/bin/python"
60
+ }
wandb/run-20220228_163348-20lja62k/files/wandb-summary.json ADDED
The diff for this file is too large to render. See raw diff
 
wandb/run-20220228_163348-20lja62k/logs/debug-internal.log ADDED
The diff for this file is too large to render. See raw diff
 
wandb/run-20220228_163348-20lja62k/logs/debug.log ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_setup.py:_flush():75] Loading settings from /home/sanchit_huggingface_co/.config/wandb/settings
2
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_setup.py:_flush():75] Loading settings from /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/settings
3
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_setup.py:_flush():75] Loading settings from environment variables: {}
4
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_setup.py:_flush():75] Inferring run settings from compute environment: {'program_relpath': 'run_speech_recognition_seq2seq.py', 'program': '/home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/run_speech_recognition_seq2seq.py'}
5
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_init.py:_log_setup():386] Logging user logs to /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_163348-20lja62k/logs/debug.log
6
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_init.py:_log_setup():387] Logging internal logs to /home/sanchit_huggingface_co/wav2vec2-gpt2-wandb-grid-search/wandb/run-20220228_163348-20lja62k/logs/debug-internal.log
7
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_init.py:init():420] calling init triggers
8
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_init.py:init():425] wandb.init called with sweep_config: {}
9
+ config: {}
10
+ 2022-02-28 16:33:48,960 INFO MainThread:228979 [wandb_init.py:init():471] starting backend
11
+ 2022-02-28 16:33:48,961 INFO MainThread:228979 [backend.py:_multiprocessing_setup():99] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
12
+ 2022-02-28 16:33:49,017 INFO MainThread:228979 [backend.py:ensure_launched():219] starting backend process...
13
+ 2022-02-28 16:33:49,073 INFO MainThread:228979 [backend.py:ensure_launched():224] started backend process with pid: 229081
14
+ 2022-02-28 16:33:49,075 INFO MainThread:228979 [wandb_init.py:init():480] backend started and connected
15
+ 2022-02-28 16:33:49,085 INFO MainThread:228979 [wandb_init.py:init():550] updated telemetry
16
+ 2022-02-28 16:33:49,224 INFO MainThread:228979 [wandb_init.py:init():581] communicating current version
17
+ 2022-02-28 16:33:49,976 INFO MainThread:228979 [wandb_init.py:init():586] got version response
18
+ 2022-02-28 16:33:49,976 INFO MainThread:228979 [wandb_init.py:init():596] communicating run to backend with 30 second timeout
19
+ 2022-02-28 16:33:50,081 INFO MainThread:228979 [wandb_init.py:init():624] starting run threads in backend
20
+ 2022-02-28 16:33:50,193 INFO MainThread:228979 [wandb_run.py:_console_start():1827] atexit reg
21
+ 2022-02-28 16:33:50,194 INFO MainThread:228979 [wandb_run.py:_redirect():1701] redirect: SettingsConsole.REDIRECT
22
+ 2022-02-28 16:33:50,194 INFO MainThread:228979 [wandb_run.py:_redirect():1706] Redirecting console.
23
+ 2022-02-28 16:33:50,196 INFO MainThread:228979 [wandb_run.py:_redirect():1762] Redirects installed.
24
+ 2022-02-28 16:33:50,196 INFO MainThread:228979 [wandb_init.py:init():651] run started, returning control to user process
25
+ 2022-02-28 16:33:50,199 INFO MainThread:228979 [wandb_run.py:_config_callback():966] config_cb None None {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': 'torch.float32', 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': False, 'is_encoder_decoder': True, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 50, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['SpeechEncoderDecoderModel'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': None, 'pad_token_id': 50256, 'eos_token_id': 50256, 'sep_token_id': None, 'decoder_start_token_id': 50256, 'task_specific_params': None, 'problem_type': None, '_name_or_path': './', 'transformers_version': None, 'decoder': {'vocab_size': 50257, 'n_positions': 1024, 'n_embd': 1024, 'n_layer': 24, 'n_head': 16, 'n_inner': None, 'activation_function': 'gelu_new', 'resid_pdrop': 0.0, 'embd_pdrop': 0.0, 'attn_pdrop': 0.0, 'layer_norm_epsilon': 1e-05, 'initializer_range': 0.02, 'summary_type': 'cls_index', 'summary_use_proj': True, 'summary_activation': None, 'summary_first_dropout': 0.0, 'summary_proj_to_labels': True, 'scale_attn_weights': True, 'use_cache': False, 'scale_attn_by_inverse_layer_idx': False, 'reorder_and_upcast_attn': False, 'bos_token_id': 50256, 'eos_token_id': 50256, 'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': True, 'cross_attention_hidden_size': None, 'add_cross_attention': True, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['GPT2LMHeadModel'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'pad_token_id': None, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': {'text-generation': {'do_sample': True, 'max_length': 50}}, 'problem_type': None, '_name_or_path': 'gpt2-medium', 'transformers_version': '4.17.0.dev0', 'n_ctx': 1024, 'n_special': 0, 'predict_special_tokens': True, 'model_type': 'gpt2'}, 'encoder': {'return_dict': True, 'output_hidden_states': False, 'output_attentions': False, 'torchscript': False, 'torch_dtype': None, 'use_bfloat16': False, 'pruned_heads': {}, 'tie_word_embeddings': True, 'is_encoder_decoder': False, 'is_decoder': False, 'cross_attention_hidden_size': None, 'add_cross_attention': False, 'tie_encoder_decoder': False, 'max_length': 20, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 1, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 1.0, 'no_repeat_ngram_size': 0, 'encoder_no_repeat_ngram_size': 0, 'bad_words_ids': None, 'num_return_sequences': 1, 'chunk_size_feed_forward': 0, 'output_scores': False, 'return_dict_in_generate': False, 'forced_bos_token_id': None, 'forced_eos_token_id': None, 'remove_invalid_values': False, 'architectures': ['Wav2Vec2ForPreTraining'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1}, 'tokenizer_class': None, 'prefix': None, 'bos_token_id': 1, 'pad_token_id': 0, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': None, 'task_specific_params': None, 'problem_type': None, '_name_or_path': 'facebook/wav2vec2-large-lv60', 'transformers_version': '4.17.0.dev0', 'feat_extract_dropout': 0.0, 'gradient_checkpointing': False, 'hidden_dropout_prob': 0.0, 'num_feat_extract_layers': 7, 'hidden_size': 1024, 'feat_extract_norm': 'layer', 'feat_extract_activation': 'gelu', 'conv_dim': [512, 512, 512, 512, 512, 512, 512], 'conv_stride': [5, 2, 2, 2, 2, 2, 2], 'conv_kernel': [10, 3, 3, 3, 3, 2, 2], 'conv_bias': True, 'num_conv_pos_embeddings': 128, 'num_conv_pos_embedding_groups': 16, 'num_hidden_layers': 24, 'intermediate_size': 4096, 'hidden_act': 'gelu', 'num_attention_heads': 16, 'hidden_dropout': 0.0, 'attention_dropout': 0.0, 'activation_dropout': 0.0, 'feat_proj_dropout': 0.0, 'final_dropout': 0.0, 'layerdrop': 0.0, 'layer_norm_eps': 1e-05, 'initializer_range': 0.02, 'vocab_size': 32, 'do_stable_layer_norm': True, 'use_weighted_layer_sum': False, 'apply_spec_augment': False, 'mask_time_prob': 0.0, 'mask_time_length': 10, 'mask_time_min_masks': 2, 'mask_feature_prob': 0.0, 'mask_feature_length': 10, 'mask_feature_min_masks': 0, 'num_codevectors_per_group': 320, 'num_codevector_groups': 2, 'contrastive_logits_temperature': 0.1, 'feat_quantizer_dropout': 0.0, 'num_negatives': 100, 'codevector_dim': 768, 'proj_codevector_dim': 768, 'diversity_loss_weight': 0.1, 'ctc_loss_reduction': 'sum', 'ctc_zero_infinity': False, 'add_adapter': True, 'adapter_kernel_size': 3, 'adapter_stride': 2, 'num_adapter_layers': 3, 'output_hidden_size': 1024, 'classifier_proj_size': 256, 'tdnn_dim': [512, 512, 512, 512, 1500], 'tdnn_kernel': [5, 3, 3, 1, 1], 'tdnn_dilation': [1, 2, 3, 1, 1], 'xvector_output_dim': 512, 'model_type': 'wav2vec2'}, 'model_type': 'speech-encoder-decoder', 'processor_class': 'Wav2Vec2Processor', 'use_cache': False, 'output_dir': './', 'overwrite_output_dir': True, 'do_train': True, 'do_eval': True, 'do_predict': False, 'evaluation_strategy': 'steps', 'prediction_loss_only': False, 'per_device_train_batch_size': 12, 'per_device_eval_batch_size': 12, 'per_gpu_train_batch_size': 'None', 'per_gpu_eval_batch_size': 'None', 'gradient_accumulation_steps': 2, 'eval_accumulation_steps': 'None', 'learning_rate': 1e-05, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 1.0, 'max_steps': -1, 'lr_scheduler_type': 'linear', 'warmup_ratio': 0.0, 'warmup_steps': 500, 'log_level': -1, 'log_level_replica': -1, 'log_on_each_node': True, 'logging_dir': './runs/Feb28_16-33-07_sanchit--v100', 'logging_strategy': 'steps', 'logging_first_step': False, 'logging_steps': 1, 'logging_nan_inf_filter': True, 'save_strategy': 'steps', 'save_steps': 500, 'save_total_limit': 1, 'save_on_each_node': False, 'no_cuda': False, 'seed': 42, 'bf16': False, 'fp16': True, 'fp16_opt_level': 'O1', 'half_precision_backend': 'amp', 'bf16_full_eval': False, 'fp16_full_eval': False, 'tf32': 'None', 'local_rank': -1, 'xpu_backend': 'None', 'tpu_num_cores': 'None', 'tpu_metrics_debug': False, 'debug': '[]', 'dataloader_drop_last': False, 'eval_steps': 500, 'dataloader_num_workers': 0, 'past_index': -1, 'run_name': './', 'disable_tqdm': False, 'remove_unused_columns': True, 'label_names': 'None', 'load_best_model_at_end': False, 'metric_for_best_model': 'None', 'greater_is_better': 'None', 'ignore_data_skip': False, 'sharded_ddp': '[]', 'deepspeed': 'None', 'label_smoothing_factor': 0.0, 'optim': 'adamw_hf', 'adafactor': False, 'group_by_length': True, 'length_column_name': 'input_length', 'report_to': "['wandb']", 'ddp_find_unused_parameters': 'None', 'ddp_bucket_cap_mb': 'None', 'dataloader_pin_memory': True, 'skip_memory_metrics': True, 'use_legacy_prediction_loop': False, 'push_to_hub': True, 'resume_from_checkpoint': 'None', 'hub_model_id': 'None', 'hub_strategy': 'every_save', 'hub_token': '<HUB_TOKEN>', 'gradient_checkpointing': True, 'fp16_backend': 'auto', 'push_to_hub_model_id': 'None', 'push_to_hub_organization': 'None', 'push_to_hub_token': '<PUSH_TO_HUB_TOKEN>', '_n_gpu': 1, 'mp_parameters': '', 'sortish_sampler': False, 'predict_with_generate': True, 'generation_max_length': 40, 'generation_num_beams': 1, 'train_batch_size': 12, 'eval_batch_size': 12}
26
+ 2022-02-28 16:33:50,201 INFO MainThread:228979 [wandb_watch.py:watch():43] Watching
wandb/run-20220228_163348-20lja62k/run-20lja62k.wandb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:33e2dea7b4b192e36f6cbf7f8bb316a85d86f1a6f9b5c53bf631cf2e61bef756
3
+ size 16238566