sanchit-gandhi HF staff commited on
Commit
a81330c
1 Parent(s): 44e7ddc

Training in progress, step 500

Browse files
.gitattributes CHANGED
@@ -27,3 +27,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
28
  wandb/run-20220322_163235-2yj5gh94/run-2yj5gh94.wandb filter=lfs diff=lfs merge=lfs -text
29
  wandb/run-20220323_165914-1vl16ira/run-1vl16ira.wandb filter=lfs diff=lfs merge=lfs -text
 
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
28
  wandb/run-20220322_163235-2yj5gh94/run-2yj5gh94.wandb filter=lfs diff=lfs merge=lfs -text
29
  wandb/run-20220323_165914-1vl16ira/run-1vl16ira.wandb filter=lfs diff=lfs merge=lfs -text
30
+ wandb/run-20220325_193848-1sz5964i/run-1sz5964i.wandb filter=lfs diff=lfs merge=lfs -text
config.json CHANGED
@@ -265,6 +265,7 @@
265
  "is_encoder_decoder": true,
266
  "max_length": 40,
267
  "model_type": "speech-encoder-decoder",
 
268
  "pad_token_id": 1,
269
  "processor_class": "Wav2Vec2Processor",
270
  "tie_word_embeddings": false,
265
  "is_encoder_decoder": true,
266
  "max_length": 40,
267
  "model_type": "speech-encoder-decoder",
268
+ "num_beams": 5,
269
  "pad_token_id": 1,
270
  "processor_class": "Wav2Vec2Processor",
271
  "tie_word_embeddings": false,
create_model.py CHANGED
@@ -13,7 +13,7 @@ model.config.decoder_start_token_id = model.decoder.config.bos_token_id
13
  model.config.pad_token_id = model.decoder.config.pad_token_id
14
  model.config.eos_token_id = model.decoder.config.eos_token_id
15
  model.config.max_length = 40
16
- model.config.num_beams = 1
17
  model.config.encoder.layerdrop = 0.0
18
  model.config.use_cache = False
19
  model.config.processor_class = "Wav2Vec2Processor"
13
  model.config.pad_token_id = model.decoder.config.pad_token_id
14
  model.config.eos_token_id = model.decoder.config.eos_token_id
15
  model.config.max_length = 40
16
+ model.config.num_beams = 5
17
  model.config.encoder.layerdrop = 0.0
18
  model.config.use_cache = False
19
  model.config.processor_class = "Wav2Vec2Processor"
pytorch_model.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c1c3887b0529e0a6a68131acba7cc1246fc8010e377d20abb6242c55ab7d7a12
3
  size 2353901617
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d8b3fb61d2800ec8274a10c55fd5f09153144a1a28327646ea158661fd675558
3
  size 2353901617
run_librispeech.sh CHANGED
@@ -1,5 +1,5 @@
1
  #!/usr/bin/env bash
2
- CUDA_VISIBLE_DEVICES=0 python run_speech_recognition_seq2seq.py \
3
  --dataset_name="librispeech_asr" \
4
  --model_name_or_path="./" \
5
  --dataset_config_name="clean" \
@@ -12,7 +12,7 @@ CUDA_VISIBLE_DEVICES=0 python run_speech_recognition_seq2seq.py \
12
  --num_train_epochs="5" \
13
  --per_device_train_batch_size="8" \
14
  --per_device_eval_batch_size="8" \
15
- --gradient_accumulation_steps="8" \
16
  --generation_max_length="40" \
17
  --generation_num_beams="1" \
18
  --learning_rate="3e-4" \
1
  #!/usr/bin/env bash
2
+ CUDA_VISIBLE_DEVICES=1 python run_speech_recognition_seq2seq.py \
3
  --dataset_name="librispeech_asr" \
4
  --model_name_or_path="./" \
5
  --dataset_config_name="clean" \
12
  --num_train_epochs="5" \
13
  --per_device_train_batch_size="8" \
14
  --per_device_eval_batch_size="8" \
15
+ --gradient_accumulation_steps="16" \
16
  --generation_max_length="40" \
17
  --generation_num_beams="1" \
18
  --learning_rate="3e-4" \
runs/Mar25_19-38-20_sanchit--v100/1648237127.7897246/events.out.tfevents.1648237127.sanchit--v100.1762967.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:405f6ecb58d922edef03c76968cce6a98ab842c25219bf61e5dd2257c71a9f4f
3
+ size 4973
runs/Mar25_19-38-20_sanchit--v100/events.out.tfevents.1648237127.sanchit--v100.1762967.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a8a0a386bb5a40454e1a6b222ab6e2318a6cf48d0e2e8da1492d14cbe38950c
3
+ size 87907
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7648918286dcb55f603f9de620875566be688909fdceaf18a2e392817ae76946
3
  size 3119
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8ab52450d2d11c0d2b8924946b765e92e1091aa39e1fe336bd8f163d51f8ae53
3
  size 3119
wandb/debug-internal.log CHANGED
@@ -1 +1 @@
1
- run-20220323_165914-1vl16ira/logs/debug-internal.log
1
+ run-20220325_193848-1sz5964i/logs/debug-internal.log
wandb/debug.log CHANGED
@@ -1 +1 @@
1
- run-20220323_165914-1vl16ira/logs/debug.log
1
+ run-20220325_193848-1sz5964i/logs/debug.log
wandb/latest-run CHANGED
@@ -1 +1 @@
1
- run-20220323_165914-1vl16ira
1
+ run-20220325_193848-1sz5964i
wandb/run-20220323_165914-1vl16ira/files/config.yaml CHANGED
@@ -8669,7 +8669,14 @@ _wandb:
8669
  - 1
8670
  - 5
8671
  - 11
 
 
 
 
 
8672
  3:
 
 
8673
  - 13
8674
  4: 3.9.5
8675
  5: 0.12.10
8669
  - 1
8670
  - 5
8671
  - 11
8672
+ 2:
8673
+ - 1
8674
+ - 5
8675
+ - 11
8676
+ - 12
8677
  3:
8678
+ - 1
8679
+ - 7
8680
  - 13
8681
  4: 3.9.5
8682
  5: 0.12.10
wandb/run-20220323_165914-1vl16ira/files/output.log CHANGED
@@ -15651,3 +15651,19 @@ Upload file wandb/run-20220323_165914-1vl16ira/run-1vl16ira.wandb: 100%|██
15651
  eval_steps_per_second = 0.593
15652
  eval_wer = 0.0924
15653
  [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
15651
  eval_steps_per_second = 0.593
15652
  eval_wer = 0.0924
15653
  [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15654
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15655
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15656
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15657
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15658
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15659
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15660
+ [INFO|trainer.py:2369] 2022-03-24 00:27:26,021 >> Batch size = 8 100%|█████████████████████████████████████████| 219M/219M [00:11<00:00, 20.8MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15661
+ 03/24/2022 00:37:28 - WARNING - huggingface_hub.repository - remote: tput: No value for $TERM and no -T specified
15662
+ remote: tput: No value for $TERM and no -T specified
15663
+ remote: tput: No value for $TERM and no -T specified
15664
+ remote: tput: No value for $TERM and no -T specified
15665
+ To https://huggingface.co/sanchit-gandhi/wav2vec2-2-bart-large-cnn
15666
+ Upload file wandb/run-20220323_165914-1vl16ira/run-1vl16ira.wandb: 100%|█████████████████████████████████████████| 220M/220M [00:11<00:00, 20.9MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15667
+ File "/home/sanchit_huggingface_co/gcp/lib/python3.9/site-packages/huggingface_hub/hf_api.py", line 870, in model_info220M [00:11<00:00, 20.9MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15668
+ File "/home/sanchit_huggingface_co/gcp/lib/python3.9/site-packages/huggingface_hub/hf_api.py", line 870, in model_info220M [00:11<00:00, 20.9MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
15669
+ File "/home/sanchit_huggingface_co/gcp/lib/python3.9/site-packages/huggingface_hub/hf_api.py", line 870, in model_info220M [00:11<00:00, 20.9MB/s]03-24 00:23:38,677 >> 5,866 >> `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...coderModel.forward` and have been ignored: input_length. If input_length are not expected by `SpeechEncoderDecoderModel.forward`, you can safely ignore this message.
wandb/run-20220323_165914-1vl16ira/files/wandb-summary.json CHANGED
The diff for this file is too large to render. See raw diff
wandb/run-20220323_165914-1vl16ira/logs/debug-internal.log CHANGED
@@ -29952,3 +29952,136 @@
29952
  2022-03-24 00:36:58,842 DEBUG SenderThread:1749633 [sender.py:send():235] send: stats
29953
  2022-03-24 00:37:13,486 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: stop_status
29954
  2022-03-24 00:37:13,486 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: stop_status
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
29952
  2022-03-24 00:36:58,842 DEBUG SenderThread:1749633 [sender.py:send():235] send: stats
29953
  2022-03-24 00:37:13,486 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: stop_status
29954
  2022-03-24 00:37:13,486 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: stop_status
29955
+ 2022-03-24 00:37:18,461 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29956
+ 2022-03-24 00:37:20,461 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29957
+ 2022-03-24 00:37:22,462 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29958
+ 2022-03-24 00:37:24,463 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29959
+ 2022-03-24 00:37:26,463 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29960
+ 2022-03-24 00:37:28,464 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29961
+ 2022-03-24 00:37:28,629 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: stop_status
29962
+ 2022-03-24 00:37:28,629 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: stop_status
29963
+ 2022-03-24 00:37:29,194 DEBUG SenderThread:1749633 [sender.py:send():235] send: stats
29964
+ 2022-03-24 00:37:29,464 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29965
+ 2022-03-24 00:37:30,465 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29966
+ 2022-03-24 00:37:34,466 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
29967
+ 2022-03-24 00:37:34,954 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
29968
+ 2022-03-24 00:37:34,954 DEBUG SenderThread:1749633 [sender.py:send():235] send: telemetry
29969
+ 2022-03-24 00:37:34,954 DEBUG SenderThread:1749633 [sender.py:send():235] send: exit
29970
+ 2022-03-24 00:37:34,955 INFO SenderThread:1749633 [sender.py:send_exit():371] handling exit code: 1
29971
+ 2022-03-24 00:37:34,955 INFO SenderThread:1749633 [sender.py:send_exit():373] handling runtime: 27499
29972
+ 2022-03-24 00:37:34,997 INFO SenderThread:1749633 [sender.py:_save_file():944] saving file wandb-summary.json with policy end
29973
+ 2022-03-24 00:37:34,998 INFO SenderThread:1749633 [sender.py:send_exit():379] send defer
29974
+ 2022-03-24 00:37:34,998 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
29975
+ 2022-03-24 00:37:34,998 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
29976
+ 2022-03-24 00:37:34,998 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 0
29977
+ 2022-03-24 00:37:34,999 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
29978
+ 2022-03-24 00:37:34,999 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 0
29979
+ 2022-03-24 00:37:34,999 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 1
29980
+ 2022-03-24 00:37:34,999 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
29981
+ 2022-03-24 00:37:34,999 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 1
29982
+ 2022-03-24 00:37:35,054 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
29983
+ 2022-03-24 00:37:35,054 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 1
29984
+ 2022-03-24 00:37:35,055 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 2
29985
+ 2022-03-24 00:37:35,055 DEBUG SenderThread:1749633 [sender.py:send():235] send: stats
29986
+ 2022-03-24 00:37:35,055 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
29987
+ 2022-03-24 00:37:35,055 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 2
29988
+ 2022-03-24 00:37:35,056 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
29989
+ 2022-03-24 00:37:35,056 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 2
29990
+ 2022-03-24 00:37:35,056 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 3
29991
+ 2022-03-24 00:37:35,056 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
29992
+ 2022-03-24 00:37:35,056 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 3
29993
+ 2022-03-24 00:37:35,104 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
29994
+ 2022-03-24 00:37:35,104 DEBUG SenderThread:1749633 [sender.py:send():235] send: summary
29995
+ 2022-03-24 00:37:35,172 INFO SenderThread:1749633 [sender.py:_save_file():944] saving file wandb-summary.json with policy end
29996
+ 2022-03-24 00:37:35,173 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
29997
+ 2022-03-24 00:37:35,173 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 3
29998
+ 2022-03-24 00:37:35,173 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 4
29999
+ 2022-03-24 00:37:35,173 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30000
+ 2022-03-24 00:37:35,173 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
30001
+ 2022-03-24 00:37:35,174 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 4
30002
+ 2022-03-24 00:37:35,174 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
30003
+ 2022-03-24 00:37:35,174 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 4
30004
+ 2022-03-24 00:37:35,275 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30005
+ 2022-03-24 00:37:35,493 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
30006
+ 2022-03-24 00:37:35,503 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/wandb-summary.json
30007
+ 2022-03-24 00:37:35,811 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 5
30008
+ 2022-03-24 00:37:35,811 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30009
+ 2022-03-24 00:37:35,811 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
30010
+ 2022-03-24 00:37:35,812 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 5
30011
+ 2022-03-24 00:37:35,812 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
30012
+ 2022-03-24 00:37:35,812 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 5
30013
+ 2022-03-24 00:37:35,812 INFO SenderThread:1749633 [dir_watcher.py:finish():283] shutting down directory watcher
30014
+ 2022-03-24 00:37:35,913 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30015
+ 2022-03-24 00:37:36,488 INFO Thread-8 :1749633 [dir_watcher.py:_on_file_modified():230] file/dir modified: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/config.yaml
30016
+ 2022-03-24 00:37:36,488 INFO SenderThread:1749633 [dir_watcher.py:finish():313] scan: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files
30017
+ 2022-03-24 00:37:36,489 INFO SenderThread:1749633 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/wandb-metadata.json wandb-metadata.json
30018
+ 2022-03-24 00:37:36,489 INFO SenderThread:1749633 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log output.log
30019
+ 2022-03-24 00:37:36,489 INFO SenderThread:1749633 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/wandb-summary.json wandb-summary.json
30020
+ 2022-03-24 00:37:36,489 INFO SenderThread:1749633 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/requirements.txt requirements.txt
30021
+ 2022-03-24 00:37:36,489 INFO SenderThread:1749633 [dir_watcher.py:finish():327] scan save: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/config.yaml config.yaml
30022
+ 2022-03-24 00:37:36,490 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 6
30023
+ 2022-03-24 00:37:36,490 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30024
+ 2022-03-24 00:37:36,491 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
30025
+ 2022-03-24 00:37:36,491 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 6
30026
+ 2022-03-24 00:37:36,493 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
30027
+ 2022-03-24 00:37:36,493 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 6
30028
+ 2022-03-24 00:37:36,494 INFO SenderThread:1749633 [file_pusher.py:finish():177] shutting down file pusher
30029
+ 2022-03-24 00:37:36,592 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30030
+ 2022-03-24 00:37:36,593 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30031
+ 2022-03-24 00:37:36,695 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30032
+ 2022-03-24 00:37:36,696 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30033
+ 2022-03-24 00:37:36,797 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30034
+ 2022-03-24 00:37:36,798 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30035
+ 2022-03-24 00:37:36,803 INFO Thread-14 :1749633 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/requirements.txt
30036
+ 2022-03-24 00:37:36,833 INFO Thread-13 :1749633 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/wandb-summary.json
30037
+ 2022-03-24 00:37:36,869 INFO Thread-12 :1749633 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/output.log
30038
+ 2022-03-24 00:37:36,899 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30039
+ 2022-03-24 00:37:36,900 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30040
+ 2022-03-24 00:37:37,001 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30041
+ 2022-03-24 00:37:37,002 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30042
+ 2022-03-24 00:37:37,103 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30043
+ 2022-03-24 00:37:37,104 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30044
+ 2022-03-24 00:37:37,205 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30045
+ 2022-03-24 00:37:37,206 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30046
+ 2022-03-24 00:37:37,307 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30047
+ 2022-03-24 00:37:37,307 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30048
+ 2022-03-24 00:37:37,409 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30049
+ 2022-03-24 00:37:37,409 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30050
+ 2022-03-24 00:37:37,511 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30051
+ 2022-03-24 00:37:37,511 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30052
+ 2022-03-24 00:37:37,517 INFO Thread-15 :1749633 [upload_job.py:push():137] Uploaded file /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/files/config.yaml
30053
+ 2022-03-24 00:37:37,612 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30054
+ 2022-03-24 00:37:37,613 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30055
+ 2022-03-24 00:37:37,714 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30056
+ 2022-03-24 00:37:37,715 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30057
+ 2022-03-24 00:37:37,717 INFO Thread-7 :1749633 [sender.py:transition_state():392] send defer: 7
30058
+ 2022-03-24 00:37:37,718 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
30059
+ 2022-03-24 00:37:37,718 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 7
30060
+ 2022-03-24 00:37:37,718 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
30061
+ 2022-03-24 00:37:37,718 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 7
30062
+ 2022-03-24 00:37:37,816 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30063
+ 2022-03-24 00:37:38,240 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 8
30064
+ 2022-03-24 00:37:38,240 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30065
+ 2022-03-24 00:37:38,241 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
30066
+ 2022-03-24 00:37:38,241 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 8
30067
+ 2022-03-24 00:37:38,241 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
30068
+ 2022-03-24 00:37:38,241 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 8
30069
+ 2022-03-24 00:37:38,241 INFO SenderThread:1749633 [sender.py:transition_state():392] send defer: 9
30070
+ 2022-03-24 00:37:38,242 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: defer
30071
+ 2022-03-24 00:37:38,243 INFO HandlerThread:1749633 [handler.py:handle_request_defer():154] handle defer: 9
30072
+ 2022-03-24 00:37:38,243 DEBUG SenderThread:1749633 [sender.py:send():235] send: final
30073
+ 2022-03-24 00:37:38,243 DEBUG SenderThread:1749633 [sender.py:send():235] send: footer
30074
+ 2022-03-24 00:37:38,243 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: defer
30075
+ 2022-03-24 00:37:38,243 INFO SenderThread:1749633 [sender.py:send_request_defer():388] handle sender defer: 9
30076
+ 2022-03-24 00:37:38,342 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: poll_exit
30077
+ 2022-03-24 00:37:38,342 DEBUG SenderThread:1749633 [sender.py:send_request():249] send_request: poll_exit
30078
+ 2022-03-24 00:37:38,342 INFO SenderThread:1749633 [file_pusher.py:join():182] waiting for file pusher
30079
+ 2022-03-24 00:37:38,403 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: get_summary
30080
+ 2022-03-24 00:37:38,485 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: sampled_history
30081
+ 2022-03-24 00:37:38,487 DEBUG HandlerThread:1749633 [handler.py:handle_request():131] handle_request: shutdown
30082
+ 2022-03-24 00:37:38,488 INFO HandlerThread:1749633 [handler.py:finish():739] shutting down handler
30083
+ 2022-03-24 00:37:39,242 INFO WriterThread:1749633 [datastore.py:close():281] close: /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220323_165914-1vl16ira/run-1vl16ira.wandb
30084
+ 2022-03-24 00:37:39,402 INFO SenderThread:1749633 [sender.py:finish():1075] shutting down sender
30085
+ 2022-03-24 00:37:39,403 INFO SenderThread:1749633 [file_pusher.py:finish():177] shutting down file pusher
30086
+ 2022-03-24 00:37:39,403 INFO SenderThread:1749633 [file_pusher.py:join():182] waiting for file pusher
30087
+ 2022-03-24 00:37:39,486 INFO MainThread:1749633 [internal.py:handle_exit():79] Internal process exited
wandb/run-20220323_165914-1vl16ira/logs/debug.log CHANGED
@@ -25,3 +25,157 @@ config: {}
25
  2022-03-23 16:59:15,529 INFO MainThread:1749532 [wandb_init.py:init():651] run started, returning control to user process
26
  2022-03-23 16:59:15,532 INFO MainThread:1749532 [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': 40, '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': 1, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': 0, 'task_specific_params': None, 'problem_type': None, '_name_or_path': './', 'transformers_version': None, 'decoder': {'vocab_size': 50264, 'max_position_embeddings': 1024, 'd_model': 1024, 'encoder_ffn_dim': 4096, 'encoder_layers': 12, 'encoder_attention_heads': 16, 'decoder_ffn_dim': 4096, 'decoder_layers': 12, 'decoder_attention_heads': 16, 'dropout': 0.1, 'attention_dropout': 0.0, 'activation_dropout': 0.0, 'activation_function': 'gelu', 'init_std': 0.02, 'encoder_layerdrop': 0.0, 'decoder_layerdrop': 0.0, 'classifier_dropout': 0.0, 'use_cache': True, 'num_hidden_layers': 12, 'scale_embedding': False, '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': 142, 'min_length': 56, 'do_sample': False, 'early_stopping': True, 'num_beams': 4, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 2.0, 'no_repeat_ngram_size': 3, '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': 0, 'forced_eos_token_id': 2, 'remove_invalid_values': False, 'architectures': ['BartForConditionalGeneration'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2}, 'tokenizer_class': None, 'prefix': ' ', 'bos_token_id': 0, 'pad_token_id': 1, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': 2, 'task_specific_params': {'summarization': {'early_stopping': True, 'length_penalty': 2.0, 'max_length': 142, 'min_length': 56, 'no_repeat_ngram_size': 3, 'num_beams': 4}}, 'problem_type': None, '_name_or_path': 'facebook/bart-large-cnn', 'transformers_version': '4.17.0.dev0', '_num_labels': 3, 'add_final_layer_norm': False, 'classif_dropout': 0.0, 'force_bos_token_to_be_generated': True, 'gradient_checkpointing': False, 'normalize_before': False, 'output_past': True, 'model_type': 'bart'}, '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': 0.1, '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': 0.1, 'attention_dropout': 0.1, 'activation_dropout': 0.1, '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': True, 'mask_time_prob': 0.1, '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': 0.0003, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 5.0, '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/Mar23_16-58-49_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': "['tensorboard', '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}
27
  2022-03-23 16:59:15,534 INFO MainThread:1749532 [wandb_watch.py:watch():43] Watching
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
25
  2022-03-23 16:59:15,529 INFO MainThread:1749532 [wandb_init.py:init():651] run started, returning control to user process
26
  2022-03-23 16:59:15,532 INFO MainThread:1749532 [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': 40, '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': 1, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': 0, 'task_specific_params': None, 'problem_type': None, '_name_or_path': './', 'transformers_version': None, 'decoder': {'vocab_size': 50264, 'max_position_embeddings': 1024, 'd_model': 1024, 'encoder_ffn_dim': 4096, 'encoder_layers': 12, 'encoder_attention_heads': 16, 'decoder_ffn_dim': 4096, 'decoder_layers': 12, 'decoder_attention_heads': 16, 'dropout': 0.1, 'attention_dropout': 0.0, 'activation_dropout': 0.0, 'activation_function': 'gelu', 'init_std': 0.02, 'encoder_layerdrop': 0.0, 'decoder_layerdrop': 0.0, 'classifier_dropout': 0.0, 'use_cache': True, 'num_hidden_layers': 12, 'scale_embedding': False, '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': 142, 'min_length': 56, 'do_sample': False, 'early_stopping': True, 'num_beams': 4, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 2.0, 'no_repeat_ngram_size': 3, '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': 0, 'forced_eos_token_id': 2, 'remove_invalid_values': False, 'architectures': ['BartForConditionalGeneration'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2}, 'tokenizer_class': None, 'prefix': ' ', 'bos_token_id': 0, 'pad_token_id': 1, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': 2, 'task_specific_params': {'summarization': {'early_stopping': True, 'length_penalty': 2.0, 'max_length': 142, 'min_length': 56, 'no_repeat_ngram_size': 3, 'num_beams': 4}}, 'problem_type': None, '_name_or_path': 'facebook/bart-large-cnn', 'transformers_version': '4.17.0.dev0', '_num_labels': 3, 'add_final_layer_norm': False, 'classif_dropout': 0.0, 'force_bos_token_to_be_generated': True, 'gradient_checkpointing': False, 'normalize_before': False, 'output_past': True, 'model_type': 'bart'}, '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': 0.1, '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': 0.1, 'attention_dropout': 0.1, 'activation_dropout': 0.1, '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': True, 'mask_time_prob': 0.1, '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': 0.0003, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 5.0, '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/Mar23_16-58-49_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': "['tensorboard', '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}
27
  2022-03-23 16:59:15,534 INFO MainThread:1749532 [wandb_watch.py:watch():43] Watching
28
+ 2022-03-24 00:37:32,662 INFO MainThread:1749532 [wandb_run.py:_atexit_cleanup():1797] got exitcode: 1
29
+ 2022-03-24 00:37:32,665 INFO MainThread:1749532 [wandb_run.py:_restore():1769] restore
30
+ 2022-03-24 00:37:34,999 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
31
+ wandb_count: 1
32
+ }
33
+ pusher_stats {
34
+ uploaded_bytes: 2015
35
+ total_bytes: 2015
36
+ }
37
+
38
+ 2022-03-24 00:37:35,174 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
39
+ wandb_count: 1
40
+ }
41
+ pusher_stats {
42
+ uploaded_bytes: 2015
43
+ total_bytes: 2015
44
+ }
45
+
46
+ 2022-03-24 00:37:35,812 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
47
+ wandb_count: 1
48
+ }
49
+ pusher_stats {
50
+ uploaded_bytes: 2015
51
+ total_bytes: 2015
52
+ }
53
+
54
+ 2022-03-24 00:37:36,491 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
55
+ wandb_count: 4
56
+ }
57
+ pusher_stats {
58
+ uploaded_bytes: 2015
59
+ total_bytes: 7427351
60
+ }
61
+
62
+ 2022-03-24 00:37:36,594 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
63
+ wandb_count: 5
64
+ }
65
+ pusher_stats {
66
+ uploaded_bytes: 2015
67
+ total_bytes: 7683909
68
+ }
69
+
70
+ 2022-03-24 00:37:36,696 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
71
+ wandb_count: 5
72
+ }
73
+ pusher_stats {
74
+ uploaded_bytes: 7427351
75
+ total_bytes: 7683909
76
+ }
77
+
78
+ 2022-03-24 00:37:36,798 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
79
+ wandb_count: 5
80
+ }
81
+ pusher_stats {
82
+ uploaded_bytes: 7427351
83
+ total_bytes: 7683909
84
+ }
85
+
86
+ 2022-03-24 00:37:36,900 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
87
+ wandb_count: 5
88
+ }
89
+ pusher_stats {
90
+ uploaded_bytes: 7427351
91
+ total_bytes: 7683909
92
+ }
93
+
94
+ 2022-03-24 00:37:37,002 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
95
+ wandb_count: 5
96
+ }
97
+ pusher_stats {
98
+ uploaded_bytes: 7427351
99
+ total_bytes: 7683909
100
+ }
101
+
102
+ 2022-03-24 00:37:37,104 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
103
+ wandb_count: 5
104
+ }
105
+ pusher_stats {
106
+ uploaded_bytes: 7427351
107
+ total_bytes: 7683909
108
+ }
109
+
110
+ 2022-03-24 00:37:37,206 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
111
+ wandb_count: 5
112
+ }
113
+ pusher_stats {
114
+ uploaded_bytes: 7427351
115
+ total_bytes: 7683909
116
+ }
117
+
118
+ 2022-03-24 00:37:37,308 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
119
+ wandb_count: 5
120
+ }
121
+ pusher_stats {
122
+ uploaded_bytes: 7427351
123
+ total_bytes: 7683909
124
+ }
125
+
126
+ 2022-03-24 00:37:37,410 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
127
+ wandb_count: 5
128
+ }
129
+ pusher_stats {
130
+ uploaded_bytes: 7683909
131
+ total_bytes: 7683909
132
+ }
133
+
134
+ 2022-03-24 00:37:37,512 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
135
+ wandb_count: 5
136
+ }
137
+ pusher_stats {
138
+ uploaded_bytes: 7683909
139
+ total_bytes: 7683909
140
+ }
141
+
142
+ 2022-03-24 00:37:37,613 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
143
+ wandb_count: 5
144
+ }
145
+ pusher_stats {
146
+ uploaded_bytes: 7683909
147
+ total_bytes: 7683909
148
+ }
149
+
150
+ 2022-03-24 00:37:37,715 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
151
+ wandb_count: 5
152
+ }
153
+ pusher_stats {
154
+ uploaded_bytes: 7683909
155
+ total_bytes: 7683909
156
+ }
157
+
158
+ 2022-03-24 00:37:38,241 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: file_counts {
159
+ wandb_count: 5
160
+ }
161
+ pusher_stats {
162
+ uploaded_bytes: 7683909
163
+ total_bytes: 7683909
164
+ }
165
+
166
+ 2022-03-24 00:37:38,403 INFO MainThread:1749532 [wandb_run.py:_wait_for_finish():1929] got exit ret: done: true
167
+ exit_result {
168
+ }
169
+ file_counts {
170
+ wandb_count: 5
171
+ }
172
+ pusher_stats {
173
+ uploaded_bytes: 7683909
174
+ total_bytes: 7683909
175
+ }
176
+ local_info {
177
+ }
178
+
179
+ 2022-03-24 00:37:39,631 INFO MainThread:1749532 [wandb_run.py:_append_history():2144] rendering history
180
+ 2022-03-24 00:37:39,631 INFO MainThread:1749532 [wandb_run.py:_append_summary():2102] rendering summary
181
+ 2022-03-24 00:37:39,632 INFO MainThread:1749532 [wandb_run.py:_append_files():2194] logging synced files
wandb/run-20220323_165914-1vl16ira/run-1vl16ira.wandb CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:8f3c44abb0b34003ae6ca342106800a262945c29dfe1f8d19ffcbf0f28c8fc34
3
- size 230211379
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:226f960b72eb3810d3a3e0c6128e1a72caef63b95117439d19faa70f8e1a011a
3
+ size 231499294
wandb/run-20220325_193848-1sz5964i/files/config.yaml ADDED
The diff for this file is too large to render. See raw diff
wandb/run-20220325_193848-1sz5964i/files/output.log ADDED
The diff for this file is too large to render. See raw diff
wandb/run-20220325_193848-1sz5964i/files/requirements.txt ADDED
@@ -0,0 +1,186 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
+ ipdb==0.13.9
53
+ ipykernel==6.8.0
54
+ ipython-genutils==0.2.0
55
+ ipython==8.0.1
56
+ ipywidgets==7.6.5
57
+ isodate==0.6.1
58
+ jax==0.2.28
59
+ jaxlib==0.1.76+cuda11.cudnn82
60
+ jedi==0.18.1
61
+ jinja2==3.0.3
62
+ jiwer==2.3.0
63
+ joblib==1.1.0
64
+ json5==0.9.6
65
+ jsonschema==4.4.0
66
+ jupyter-client==7.1.2
67
+ jupyter-console==6.4.0
68
+ jupyter-core==4.9.1
69
+ jupyter-server==1.13.5
70
+ jupyter==1.0.0
71
+ jupyterlab-pygments==0.1.2
72
+ jupyterlab-server==2.10.3
73
+ jupyterlab-widgets==1.0.2
74
+ jupyterlab==3.2.9
75
+ kiwisolver==1.3.2
76
+ librosa==0.8.1
77
+ llvmlite==0.38.0
78
+ markdown==3.3.6
79
+ markupsafe==2.0.1
80
+ matplotlib-inline==0.1.3
81
+ matplotlib==3.5.1
82
+ mistune==0.8.4
83
+ msgpack==1.0.3
84
+ multidict==6.0.2
85
+ multiprocess==0.70.12.2
86
+ mypy-extensions==0.4.3
87
+ nbclassic==0.3.5
88
+ nbclient==0.5.10
89
+ nbconvert==6.4.1
90
+ nbformat==5.1.3
91
+ nest-asyncio==1.5.4
92
+ notebook==6.4.8
93
+ numba==0.55.1
94
+ numpy==1.21.5
95
+ oauthlib==3.2.0
96
+ opt-einsum==3.3.0
97
+ optax==0.1.0
98
+ packaging==21.3
99
+ pandas==1.4.0
100
+ pandocfilters==1.5.0
101
+ parso==0.8.3
102
+ pathspec==0.9.0
103
+ pathtools==0.1.2
104
+ pexpect==4.8.0
105
+ phonemizer==3.0.1
106
+ pickleshare==0.7.5
107
+ pillow==9.0.0
108
+ pip==22.0.2
109
+ pkg-resources==0.0.0
110
+ platformdirs==2.4.1
111
+ pooch==1.6.0
112
+ prometheus-client==0.13.1
113
+ promise==2.3
114
+ prompt-toolkit==3.0.26
115
+ protobuf==3.19.4
116
+ psutil==5.9.0
117
+ ptyprocess==0.7.0
118
+ pure-eval==0.2.2
119
+ pyarrow==6.0.1
120
+ pyasn1-modules==0.2.8
121
+ pyasn1==0.4.8
122
+ pycparser==2.21
123
+ pyctcdecode==0.3.0
124
+ pygments==2.11.2
125
+ pygtrie==2.4.2
126
+ pyparsing==3.0.7
127
+ pyrsistent==0.18.1
128
+ python-dateutil==2.8.2
129
+ python-levenshtein==0.12.2
130
+ pytz==2021.3
131
+ pyyaml==6.0
132
+ pyzmq==22.3.0
133
+ qtconsole==5.2.2
134
+ qtpy==2.0.1
135
+ regex==2022.1.18
136
+ requests-oauthlib==1.3.1
137
+ requests==2.27.1
138
+ resampy==0.2.2
139
+ rfc3986==2.0.0
140
+ rsa==4.8
141
+ sacremoses==0.0.47
142
+ scikit-learn==1.0.2
143
+ scipy==1.7.3
144
+ segments==2.2.0
145
+ send2trash==1.8.0
146
+ sentry-sdk==1.5.6
147
+ setuptools==44.1.1
148
+ shortuuid==1.0.8
149
+ six==1.16.0
150
+ smmap==5.0.0
151
+ sniffio==1.2.0
152
+ sortedcontainers==2.4.0
153
+ soundfile==0.10.3.post1
154
+ stack-data==0.1.4
155
+ tabulate==0.8.9
156
+ tensorboard-data-server==0.6.1
157
+ tensorboard-plugin-wit==1.8.1
158
+ tensorboard==2.8.0
159
+ termcolor==1.1.0
160
+ terminado==0.13.1
161
+ testpath==0.5.0
162
+ threadpoolctl==3.1.0
163
+ tokenizers==0.11.4
164
+ toml==0.10.2
165
+ tomli==2.0.0
166
+ toolz==0.11.2
167
+ torch==1.10.2+cu113
168
+ torchaudio==0.10.2+cu113
169
+ tornado==6.1
170
+ tqdm==4.62.3
171
+ traitlets==5.1.1
172
+ transformers==4.17.0.dev0
173
+ typing-extensions==3.10.0.2
174
+ uritemplate==4.1.1
175
+ urllib3==1.26.8
176
+ wandb==0.12.10
177
+ wcwidth==0.2.5
178
+ webencodings==0.5.1
179
+ websocket-client==1.2.3
180
+ werkzeug==2.0.2
181
+ wheel==0.37.1
182
+ widgetsnbextension==3.5.2
183
+ xxhash==2.0.2
184
+ yarl==1.7.2
185
+ yaspin==2.1.0
186
+ zipp==3.7.0
wandb/run-20220325_193848-1sz5964i/files/wandb-metadata.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "os": "Linux-5.11.0-1028-gcp-x86_64-with-glibc2.33",
3
+ "python": "3.9.5",
4
+ "heartbeatAt": "2022-03-25T19:38:49.160439",
5
+ "startedAt": "2022-03-25T19:38:48.113785",
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
+ "--dataset_config_name=clean",
15
+ "--train_split_name=train.100",
16
+ "--eval_split_name=validation",
17
+ "--output_dir=./",
18
+ "--preprocessing_num_workers=1",
19
+ "--length_column_name=input_length",
20
+ "--overwrite_output_dir",
21
+ "--num_train_epochs=5",
22
+ "--per_device_train_batch_size=8",
23
+ "--per_device_eval_batch_size=8",
24
+ "--gradient_accumulation_steps=16",
25
+ "--generation_max_length=40",
26
+ "--generation_num_beams=1",
27
+ "--learning_rate=3e-4",
28
+ "--warmup_steps=500",
29
+ "--evaluation_strategy=steps",
30
+ "--text_column_name=text",
31
+ "--save_steps=500",
32
+ "--eval_steps=500",
33
+ "--logging_steps=1",
34
+ "--save_total_limit=1",
35
+ "--freeze_feature_encoder",
36
+ "--gradient_checkpointing",
37
+ "--fp16",
38
+ "--group_by_length",
39
+ "--predict_with_generate",
40
+ "--do_lower_case",
41
+ "--do_eval",
42
+ "--do_train",
43
+ "--push_to_hub",
44
+ "--use_auth_token"
45
+ ],
46
+ "state": "running",
47
+ "program": "/home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/run_speech_recognition_seq2seq.py",
48
+ "codePath": "run_speech_recognition_seq2seq.py",
49
+ "git": {
50
+ "remote": "https://huggingface.co/sanchit-gandhi/wav2vec2-2-bart-large-cnn",
51
+ "commit": "44e7ddc68f58e14a4c365cc3f7fd5b4ece1f1615"
52
+ },
53
+ "email": "sanchit@huggingface.co",
54
+ "root": "/home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn",
55
+ "host": "sanchit--v100",
56
+ "username": "sanchit_huggingface_co",
57
+ "executable": "/home/sanchit_huggingface_co/gcp/bin/python"
58
+ }
wandb/run-20220325_193848-1sz5964i/files/wandb-summary.json ADDED
The diff for this file is too large to render. See raw diff
wandb/run-20220325_193848-1sz5964i/logs/debug-internal.log ADDED
The diff for this file is too large to render. See raw diff
wandb/run-20220325_193848-1sz5964i/logs/debug.log ADDED
@@ -0,0 +1,27 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_setup.py:_flush():75] Loading settings from /home/sanchit_huggingface_co/.config/wandb/settings
2
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_setup.py:_flush():75] Loading settings from /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/settings
3
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_setup.py:_flush():75] Loading settings from environment variables: {}
4
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_setup.py:_flush():75] Inferring run settings from compute environment: {'program_relpath': 'run_speech_recognition_seq2seq.py', 'program': '/home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/run_speech_recognition_seq2seq.py'}
5
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_init.py:_log_setup():386] Logging user logs to /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220325_193848-1sz5964i/logs/debug.log
6
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_init.py:_log_setup():387] Logging internal logs to /home/sanchit_huggingface_co/wav2vec2-2-bart-large-cnn/wandb/run-20220325_193848-1sz5964i/logs/debug-internal.log
7
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_init.py:init():420] calling init triggers
8
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_init.py:init():425] wandb.init called with sweep_config: {}
9
+ config: {}
10
+ 2022-03-25 19:38:48,115 INFO MainThread:1762967 [wandb_init.py:init():471] starting backend
11
+ 2022-03-25 19:38:48,116 INFO MainThread:1762967 [backend.py:_multiprocessing_setup():99] multiprocessing start_methods=fork,spawn,forkserver, using: spawn
12
+ 2022-03-25 19:38:48,169 INFO MainThread:1762967 [backend.py:ensure_launched():219] starting backend process...
13
+ 2022-03-25 19:38:48,239 INFO MainThread:1762967 [backend.py:ensure_launched():224] started backend process with pid: 1763281
14
+ 2022-03-25 19:38:48,242 INFO MainThread:1762967 [wandb_init.py:init():480] backend started and connected
15
+ 2022-03-25 19:38:48,253 INFO MainThread:1762967 [wandb_init.py:init():550] updated telemetry
16
+ 2022-03-25 19:38:48,370 INFO MainThread:1762967 [wandb_init.py:init():581] communicating current version
17
+ 2022-03-25 19:38:49,050 INFO MainThread:1762967 [wandb_init.py:init():586] got version response upgrade_message: "wandb version 0.12.11 is available! To upgrade, please run:\n $ pip install wandb --upgrade"
18
+
19
+ 2022-03-25 19:38:49,051 INFO MainThread:1762967 [wandb_init.py:init():596] communicating run to backend with 30 second timeout
20
+ 2022-03-25 19:38:49,155 INFO MainThread:1762967 [wandb_init.py:init():624] starting run threads in backend
21
+ 2022-03-25 19:38:49,257 INFO MainThread:1762967 [wandb_run.py:_console_start():1827] atexit reg
22
+ 2022-03-25 19:38:49,258 INFO MainThread:1762967 [wandb_run.py:_redirect():1701] redirect: SettingsConsole.REDIRECT
23
+ 2022-03-25 19:38:49,258 INFO MainThread:1762967 [wandb_run.py:_redirect():1706] Redirecting console.
24
+ 2022-03-25 19:38:49,259 INFO MainThread:1762967 [wandb_run.py:_redirect():1762] Redirects installed.
25
+ 2022-03-25 19:38:49,259 INFO MainThread:1762967 [wandb_init.py:init():651] run started, returning control to user process
26
+ 2022-03-25 19:38:49,262 INFO MainThread:1762967 [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': 40, 'min_length': 0, 'do_sample': False, 'early_stopping': False, 'num_beams': 5, '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': 1, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': 0, 'task_specific_params': None, 'problem_type': None, '_name_or_path': './', 'transformers_version': None, 'decoder': {'vocab_size': 50264, 'max_position_embeddings': 1024, 'd_model': 1024, 'encoder_ffn_dim': 4096, 'encoder_layers': 12, 'encoder_attention_heads': 16, 'decoder_ffn_dim': 4096, 'decoder_layers': 12, 'decoder_attention_heads': 16, 'dropout': 0.1, 'attention_dropout': 0.0, 'activation_dropout': 0.0, 'activation_function': 'gelu', 'init_std': 0.02, 'encoder_layerdrop': 0.0, 'decoder_layerdrop': 0.0, 'classifier_dropout': 0.0, 'use_cache': True, 'num_hidden_layers': 12, 'scale_embedding': False, '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': 142, 'min_length': 56, 'do_sample': False, 'early_stopping': True, 'num_beams': 4, 'num_beam_groups': 1, 'diversity_penalty': 0.0, 'temperature': 1.0, 'top_k': 50, 'top_p': 1.0, 'repetition_penalty': 1.0, 'length_penalty': 2.0, 'no_repeat_ngram_size': 3, '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': 0, 'forced_eos_token_id': 2, 'remove_invalid_values': False, 'architectures': ['BartForConditionalGeneration'], 'finetuning_task': None, 'id2label': {0: 'LABEL_0', 1: 'LABEL_1', 2: 'LABEL_2'}, 'label2id': {'LABEL_0': 0, 'LABEL_1': 1, 'LABEL_2': 2}, 'tokenizer_class': None, 'prefix': ' ', 'bos_token_id': 0, 'pad_token_id': 1, 'eos_token_id': 2, 'sep_token_id': None, 'decoder_start_token_id': 2, 'task_specific_params': {'summarization': {'early_stopping': True, 'length_penalty': 2.0, 'max_length': 142, 'min_length': 56, 'no_repeat_ngram_size': 3, 'num_beams': 4}}, 'problem_type': None, '_name_or_path': 'facebook/bart-large-cnn', 'transformers_version': '4.17.0.dev0', '_num_labels': 3, 'add_final_layer_norm': False, 'classif_dropout': 0.0, 'force_bos_token_to_be_generated': True, 'gradient_checkpointing': False, 'normalize_before': False, 'output_past': True, 'model_type': 'bart'}, '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': 0.1, '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': 0.1, 'attention_dropout': 0.1, 'activation_dropout': 0.1, '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': True, 'mask_time_prob': 0.1, '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': 16, 'eval_accumulation_steps': 'None', 'learning_rate': 0.0003, 'weight_decay': 0.0, 'adam_beta1': 0.9, 'adam_beta2': 0.999, 'adam_epsilon': 1e-08, 'max_grad_norm': 1.0, 'num_train_epochs': 5.0, '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/Mar25_19-38-20_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': "['tensorboard', '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}
27
+ 2022-03-25 19:38:49,265 INFO MainThread:1762967 [wandb_watch.py:watch():43] Watching
wandb/run-20220325_193848-1sz5964i/run-1sz5964i.wandb ADDED
@@ -0,0 +1,3 @@
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67e73417410b53b49b6c9ce3ea733121b35047d03374b594e6a0e5a83822a8a4
3
+ size 100003253