Hugging Face's logo

arielcerdap
/
whisper-largev3turbo-fluencybank

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
English
whisper
stuttered-speech
speech-recognition
asr
disfluency
fluencybank
Generated from Trainer
Eval Results (legacy)
Model card Files
xet
 Metrics Community
arielcerdap commited on
Commit
3b27719
·
verified ·
1 Parent(s): 1d5a214

End of training

Browse files
Files changed (7) hide show
  1. README.md +20 -27
  2. logs/events.out.tfevents.1774442780.515c0c429b4d.399.0 +3 -0
  3. logs/events.out.tfevents.1774442829.515c0c429b4d.399.1 +3 -0
  4. logs/events.out.tfevents.1774442849.515c0c429b4d.399.2 +3 -0
  5. logs/events.out.tfevents.1774442862.515c0c429b4d.399.3 +3 -0
  6. model.safetensors +1 -1
  7. training_args.bin +1 -1
README.md CHANGED
@@ -29,7 +29,7 @@ model-index:
29
  metrics:
30
  - name: Wer
31
  type: wer
32
- value: 13.711272247857615
33
  ---
34
 
35
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -39,9 +39,9 @@ should probably proofread and complete it, then remove this comment. -->
39
 
40
  This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the FluencyBank Timestamped dataset.
41
  It achieves the following results on the evaluation set:
42
- - Loss: 1.9454
43
- - Wer: 13.7113
44
- - Cer: 8.2935
45
 
46
  ## Model description
47
 
@@ -61,38 +61,31 @@ More information needed
61
 
62
  The following hyperparameters were used during training:
63
  - learning_rate: 8e-06
64
- - train_batch_size: 8
65
- - eval_batch_size: 8
66
  - seed: 42
67
  - gradient_accumulation_steps: 4
68
- - total_train_batch_size: 32
69
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
70
  - lr_scheduler_type: cosine
71
  - lr_scheduler_warmup_ratio: 0.1
72
- - training_steps: 8000
73
- - mixed_precision_training: Native AMP
74
  - label_smoothing_factor: 0.1
75
 
76
  ### Training results
77
 
78
- | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
79
- |:-------------:|:-------:|:----:|:---------------:|:-------:|:-------:|
80
- | 1.5043 | 5.8480 | 500 | 1.6909 | 11.8655 | 6.0676 |
81
- | 1.4553 | 11.6959 | 1000 | 1.7112 | 20.5889 | 15.3853 |
82
- | 1.4374 | 17.5439 | 1500 | 1.7597 | 10.8987 | 6.2087 |
83
- | 1.4332 | 23.3918 | 2000 | 1.7736 | 11.9974 | 7.3968 |
84
- | 1.4265 | 29.2398 | 2500 | 1.8074 | 12.5027 | 7.5515 |
85
- | 1.4202 | 35.0877 | 3000 | 1.8468 | 11.9095 | 7.1737 |
86
- | 1.4182 | 40.9357 | 3500 | 1.8376 | 12.0633 | 7.5379 |
87
- | 1.4168 | 46.7836 | 4000 | 1.8712 | 11.4480 | 7.1692 |
88
- | 1.4156 | 52.6316 | 4500 | 1.8798 | 11.2063 | 6.7459 |
89
- | 1.4153 | 58.4795 | 5000 | 1.9020 | 12.2610 | 7.4787 |
90
- | 1.4145 | 64.3275 | 5500 | 1.9188 | 12.8104 | 7.9385 |
91
- | 1.4142 | 70.1754 | 6000 | 1.9312 | 13.1839 | 8.1478 |
92
- | 1.4141 | 76.0234 | 6500 | 1.9387 | 13.4696 | 8.1569 |
93
- | 1.4141 | 81.8713 | 7000 | 1.9431 | 13.6893 | 8.2981 |
94
- | 1.414 | 87.7193 | 7500 | 1.9451 | 13.7113 | 8.2935 |
95
- | 1.414 | 93.5673 | 8000 | 1.9454 | 13.7113 | 8.2935 |
96
 
97
 
98
  ### Framework versions
 
29
  metrics:
30
  - name: Wer
31
  type: wer
32
+ value: 9.97582948802461
33
  ---
34
 
35
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
39
 
40
  This model is a fine-tuned version of [openai/whisper-large-v3-turbo](https://huggingface.co/openai/whisper-large-v3-turbo) on the FluencyBank Timestamped dataset.
41
  It achieves the following results on the evaluation set:
42
+ - Loss: 1.8510
43
+ - Wer: 9.9758
44
+ - Cer: 5.8628
45
 
46
  ## Model description
47
 
 
61
 
62
  The following hyperparameters were used during training:
63
  - learning_rate: 8e-06
64
+ - train_batch_size: 32
65
+ - eval_batch_size: 32
66
  - seed: 42
67
  - gradient_accumulation_steps: 4
68
+ - total_train_batch_size: 128
69
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
70
  - lr_scheduler_type: cosine
71
  - lr_scheduler_warmup_ratio: 0.1
72
+ - training_steps: 2500
 
73
  - label_smoothing_factor: 0.1
74
 
75
  ### Training results
76
 
77
+ | Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
78
+ |:-------------:|:--------:|:----:|:---------------:|:-------:|:------:|
79
+ | 1.4697 | 11.6279 | 250 | 1.7083 | 12.7005 | 6.4682 |
80
+ | 1.4366 | 23.2558 | 500 | 1.7474 | 9.8879 | 5.6989 |
81
+ | 1.4266 | 34.8837 | 750 | 1.7590 | 9.9978 | 5.8992 |
82
+ | 1.4248 | 46.5116 | 1000 | 1.7597 | 10.2395 | 6.0267 |
83
+ | 1.4195 | 58.1395 | 1250 | 1.8063 | 9.6902 | 5.6580 |
84
+ | 1.4183 | 69.7674 | 1500 | 1.8249 | 9.9978 | 5.8492 |
85
+ | 1.4176 | 81.3953 | 1750 | 1.8405 | 9.9319 | 5.8309 |
86
+ | 1.4173 | 93.0233 | 2000 | 1.8477 | 9.9978 | 5.8901 |
87
+ | 1.4172 | 104.6512 | 2250 | 1.8514 | 9.9539 | 5.8810 |
88
+ | 1.4172 | 116.2791 | 2500 | 1.8510 | 9.9758 | 5.8628 |
 
 
 
 
 
 
89
 
90
 
91
  ### Framework versions
logs/events.out.tfevents.1774442780.515c0c429b4d.399.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:779489db6e7ea70d7740bc4a9a92045011ff575a72ad08743f02ed6cece29c63
3
+ size 5824
logs/events.out.tfevents.1774442829.515c0c429b4d.399.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e7395ed0de5fc2bb948f5426fbef41ecd03a7bd37d632b6047e039709d29719
3
+ size 5823
logs/events.out.tfevents.1774442849.515c0c429b4d.399.2 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:23cc5b4c967de7447d8427277bf88a4c3da1b4d4aea5d80b5e5ece5d8e67e350
3
+ size 5823
logs/events.out.tfevents.1774442862.515c0c429b4d.399.3 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cd5385f46d72c6cd8a4a19f3d5a24107a2ce6920960c36c9371a1817ffc520c0
3
+ size 20369
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b5c8ba1fea3c18a48160b27b1b1480135842b02bcd5caa065b203ad8d8f81ead
3
  size 3235581408
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5d765f48eb83d8b6b2ae11fd87742cd18e917cdb33678d4c9eb12bb7f42b8524
3
  size 3235581408
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:0c9f5a858e954ad78efaff51de62e058a1b7df66dd5e8aedfde8152362f1b098
3
  size 5841
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:88383f8a1c783b2beb92901bbf8b691425468adb0551896a646f0e9a3fa17c20
3
  size 5841