Hugging Face's logo

NbAiLab
/
whisper-tiny-gputest

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
whisper
whisper-event
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
 Metrics Community
2
pere commited on
Commit
1189f37
·
1 Parent(s): e707816

Training in progress, step 200

Browse files
Files changed (18) hide show
  1. .gitignore +1 -0
  2. added_tokens.json +1 -2
  3. config.json +10 -1
  4. merges.txt +1 -1
  5. preprocessor_config.json +0 -0
  6. pytorch_model.bin +3 -0
  7. run.sh +0 -1
  8. runs/Mar17_11-26-25_dante/1679048805.5259922/events.out.tfevents.1679048805.dante.3289286.1 +3 -0
  9. runs/Mar17_11-26-25_dante/events.out.tfevents.1679048805.dante.3289286.0 +3 -0
  10. runs/Mar17_11-44-55_dante/1679049913.0870838/events.out.tfevents.1679049913.dante.3379179.1 +3 -0
  11. runs/Mar17_11-44-55_dante/events.out.tfevents.1679049913.dante.3379179.0 +3 -0
  12. runs/Mar17_11-46-06_dante/1679049983.539156/events.out.tfevents.1679049983.dante.3389725.1 +3 -0
  13. runs/Mar17_11-46-06_dante/events.out.tfevents.1679049983.dante.3389725.0 +3 -0
  14. special_tokens_map.json +2 -2
  15. tokenizer.json +0 -0
  16. tokenizer_config.json +1 -2
  17. training_args.bin +2 -2
  18. vocab.json +0 -0
.gitignore ADDED
@@ -0,0 +1 @@
 
 
1
+ checkpoint-*/
added_tokens.json CHANGED
@@ -17,7 +17,6 @@
17
  "<|da|>": 50285,
18
  "<|de|>": 50261,
19
  "<|el|>": 50281,
20
- "<|endoftext|>": 50257,
21
  "<|en|>": 50259,
22
  "<|es|>": 50262,
23
  "<|et|>": 50307,
@@ -30,6 +29,7 @@
30
  "<|gu|>": 50333,
31
  "<|haw|>": 50352,
32
  "<|ha|>": 50354,
 
33
  "<|hi|>": 50276,
34
  "<|hr|>": 50291,
35
  "<|ht|>": 50339,
@@ -38,7 +38,6 @@
38
  "<|id|>": 50275,
39
  "<|is|>": 50311,
40
  "<|it|>": 50274,
41
- "<|iw|>": 50279,
42
  "<|ja|>": 50266,
43
  "<|jw|>": 50356,
44
  "<|ka|>": 50329,
 
17
  "<|da|>": 50285,
18
  "<|de|>": 50261,
19
  "<|el|>": 50281,
 
20
  "<|en|>": 50259,
21
  "<|es|>": 50262,
22
  "<|et|>": 50307,
 
29
  "<|gu|>": 50333,
30
  "<|haw|>": 50352,
31
  "<|ha|>": 50354,
32
+ "<|he|>": 50279,
33
  "<|hi|>": 50276,
34
  "<|hr|>": 50291,
35
  "<|ht|>": 50339,
 
38
  "<|id|>": 50275,
39
  "<|is|>": 50311,
40
  "<|it|>": 50274,
 
41
  "<|ja|>": 50266,
42
  "<|jw|>": 50356,
43
  "<|ka|>": 50329,
config.json CHANGED
@@ -2,6 +2,7 @@
2
  "_name_or_path": "openai/whisper-tiny",
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
 
5
  "architectures": [
6
  "WhisperForConditionalGeneration"
7
  ],
@@ -11,6 +12,7 @@
11
  50257
12
  ],
13
  "bos_token_id": 50257,
 
14
  "d_model": 384,
15
  "decoder_attention_heads": 6,
16
  "decoder_ffn_dim": 1536,
@@ -26,6 +28,12 @@
26
  "forced_decoder_ids": null,
27
  "init_std": 0.02,
28
  "is_encoder_decoder": true,
 
 
 
 
 
 
29
  "max_length": 448,
30
  "max_source_positions": 1500,
31
  "max_target_positions": 448,
@@ -35,7 +43,8 @@
35
  "pad_token_id": 50257,
36
  "scale_embedding": false,
37
  "torch_dtype": "float32",
38
- "transformers_version": "4.26.0.dev0",
39
  "use_cache": false,
 
40
  "vocab_size": 51865
41
  }
 
2
  "_name_or_path": "openai/whisper-tiny",
3
  "activation_dropout": 0.0,
4
  "activation_function": "gelu",
5
+ "apply_spec_augment": false,
6
  "architectures": [
7
  "WhisperForConditionalGeneration"
8
  ],
 
12
  50257
13
  ],
14
  "bos_token_id": 50257,
15
+ "classifier_proj_size": 256,
16
  "d_model": 384,
17
  "decoder_attention_heads": 6,
18
  "decoder_ffn_dim": 1536,
 
28
  "forced_decoder_ids": null,
29
  "init_std": 0.02,
30
  "is_encoder_decoder": true,
31
+ "mask_feature_length": 10,
32
+ "mask_feature_min_masks": 0,
33
+ "mask_feature_prob": 0.0,
34
+ "mask_time_length": 10,
35
+ "mask_time_min_masks": 2,
36
+ "mask_time_prob": 0.05,
37
  "max_length": 448,
38
  "max_source_positions": 1500,
39
  "max_target_positions": 448,
 
43
  "pad_token_id": 50257,
44
  "scale_embedding": false,
45
  "torch_dtype": "float32",
46
+ "transformers_version": "4.28.0.dev0",
47
  "use_cache": false,
48
+ "use_weighted_layer_sum": false,
49
  "vocab_size": 51865
50
  }
merges.txt CHANGED
@@ -1,4 +1,4 @@
1
- #version: 0.2
2
  Ġ a
3
  Ġt h
4
  i n
 
1
+ #version: 0.2 - Trained by `huggingface/tokenizers`
2
  Ġ a
3
  Ġt h
4
  i n
preprocessor_config.json CHANGED
The diff for this file is too large to render. See raw diff
 
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:12553e0e927eed2c74d99b6cf8b63f39c52c3d7ddf6685b691ef4eaf737aa78b
3
+ size 151098921
run.sh CHANGED
@@ -9,7 +9,6 @@ python run_speech_recognition_seq2seq_streaming.py \
9
  --output_dir="./" \
10
  --per_device_train_batch_size="128" \
11
  --per_device_eval_batch_size="32" \
12
- --gradient_accumulation_step="1" \
13
  --logging_steps="200" \
14
  --learning_rate="3e-6" \
15
  --lr_scheduler_type="constant_with_warmup" \
 
9
  --output_dir="./" \
10
  --per_device_train_batch_size="128" \
11
  --per_device_eval_batch_size="32" \
 
12
  --logging_steps="200" \
13
  --learning_rate="3e-6" \
14
  --lr_scheduler_type="constant_with_warmup" \
runs/Mar17_11-26-25_dante/1679048805.5259922/events.out.tfevents.1679048805.dante.3289286.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb323c0adaf46f08789eaa656ce29f54a414144a4d7089e6e5e6b6b78d816386
3
+ size 6000
runs/Mar17_11-26-25_dante/events.out.tfevents.1679048805.dante.3289286.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:96111a5cc0a61b189c72c37afde9e60957029a82678ef276b338089b3ad5975d
3
+ size 4657
runs/Mar17_11-44-55_dante/1679049913.0870838/events.out.tfevents.1679049913.dante.3379179.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dad130fea0ce6160e7e3f977432f0dbe443d9e7b5528bc3186c98b110a5a70d4
3
+ size 6000
runs/Mar17_11-44-55_dante/events.out.tfevents.1679049913.dante.3379179.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c903331b7bf4b5917942facf549c8fd588774ef29d1c1eb5ff53e2c52c70337
3
+ size 4136
runs/Mar17_11-46-06_dante/1679049983.539156/events.out.tfevents.1679049983.dante.3389725.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:529285418f7d91b1c6a78c8472f127c1d56cde29c73e72b738f8f57183c2a5bf
3
+ size 6000
runs/Mar17_11-46-06_dante/events.out.tfevents.1679049983.dante.3389725.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f68fc2531410ac9a6101deeb5352ac6cd5136feb1657011faea65415aa401bab
3
+ size 5132
special_tokens_map.json CHANGED
@@ -22,7 +22,7 @@
22
  "<|hi|>",
23
  "<|fi|>",
24
  "<|vi|>",
25
- "<|iw|>",
26
  "<|uk|>",
27
  "<|el|>",
28
  "<|ms|>",
@@ -124,7 +124,7 @@
124
  },
125
  "pad_token": "<|endoftext|>",
126
  "unk_token": {
127
- "content": "",
128
  "lstrip": false,
129
  "normalized": true,
130
  "rstrip": false,
 
22
  "<|hi|>",
23
  "<|fi|>",
24
  "<|vi|>",
25
+ "<|he|>",
26
  "<|uk|>",
27
  "<|el|>",
28
  "<|ms|>",
 
124
  },
125
  "pad_token": "<|endoftext|>",
126
  "unk_token": {
127
+ "content": "<|endoftext|>",
128
  "lstrip": false,
129
  "normalized": true,
130
  "rstrip": false,
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json CHANGED
@@ -19,7 +19,6 @@
19
  },
20
  "errors": "replace",
21
  "model_max_length": 1024,
22
- "name_or_path": "openai/whisper-tiny",
23
  "pad_token": null,
24
  "processor_class": "WhisperProcessor",
25
  "return_attention_mask": false,
@@ -27,7 +26,7 @@
27
  "tokenizer_class": "WhisperTokenizer",
28
  "unk_token": {
29
  "__type": "AddedToken",
30
- "content": "",
31
  "lstrip": false,
32
  "normalized": true,
33
  "rstrip": false,
 
19
  },
20
  "errors": "replace",
21
  "model_max_length": 1024,
 
22
  "pad_token": null,
23
  "processor_class": "WhisperProcessor",
24
  "return_attention_mask": false,
 
26
  "tokenizer_class": "WhisperTokenizer",
27
  "unk_token": {
28
  "__type": "AddedToken",
29
+ "content": "<|endoftext|>",
30
  "lstrip": false,
31
  "normalized": true,
32
  "rstrip": false,
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:402de2c4bd980e7aa9dfbd8a32ec503e2002f70e226ac0ad0b28d4f63153cc68
3
- size 3643
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2e571ee67bac8da76cc9d321b972299d9ace7203f12f58e3e1067698c4cb42d0
3
+ size 3707
vocab.json CHANGED
The diff for this file is too large to render. See raw diff