Hugging Face's logo

sujithvemi
/
whisper-medium-physician-dictation-gpt-4-turbo

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
English
whisper
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
sujithvemi commited on
Commit
120bad2
•
1 Parent(s): 0f969a5

End of training

Browse files
Files changed (2) hide show
  1. README.md +20 -21
  2. model.safetensors +1 -1
README.md CHANGED
@@ -24,7 +24,7 @@ model-index:
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 5.2149306023446975
28
  ---
29
 
30
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -34,8 +34,8 @@ should probably proofread and complete it, then remove this comment. -->
34
 
35
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Physician Dictation GPT 4 Turbo dataset.
36
  It achieves the following results on the evaluation set:
37
- - Loss: 0.1587
38
- - Wer: 5.2149
39
 
40
  ## Model description
41
 
@@ -60,7 +60,6 @@ The following hyperparameters were used during training:
60
  - seed: 42
61
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
62
  - lr_scheduler_type: linear
63
- - lr_scheduler_warmup_steps: 500
64
  - training_steps: 8500
65
  - mixed_precision_training: Native AMP
66
 
@@ -68,23 +67,23 @@ The following hyperparameters were used during training:
68
 
69
  | Training Loss | Epoch | Step | Validation Loss | Wer |
70
  |:-------------:|:-------:|:----:|:---------------:|:------:|
71
- | 0.0314 | 3.9683 | 500 | 0.1036 | 4.6175 |
72
- | 0.0088 | 7.9365 | 1000 | 0.1090 | 5.0712 |
73
- | 0.0009 | 11.9048 | 1500 | 0.1229 | 5.1880 |
74
- | 0.0013 | 15.8730 | 2000 | 0.1329 | 5.6506 |
75
- | 0.0031 | 19.8413 | 2500 | 0.1360 | 5.3227 |
76
- | 0.0004 | 23.8095 | 3000 | 0.1288 | 4.9589 |
77
- | 0.0008 | 27.7778 | 3500 | 0.1381 | 4.9724 |
78
- | 0.0001 | 31.7460 | 4000 | 0.1391 | 5.0667 |
79
- | 0.0001 | 35.7143 | 4500 | 0.1435 | 5.4710 |
80
- | 0.0001 | 39.6825 | 5000 | 0.1445 | 5.2688 |
81
- | 0.0002 | 43.6508 | 5500 | 0.1490 | 5.1880 |
82
- | 0.0 | 47.6190 | 6000 | 0.1510 | 5.3093 |
83
- | 0.0 | 51.5873 | 6500 | 0.1540 | 5.2868 |
84
- | 0.0 | 55.5556 | 7000 | 0.1559 | 5.3452 |
85
- | 0.0 | 59.5238 | 7500 | 0.1573 | 5.2104 |
86
- | 0.0 | 63.4921 | 8000 | 0.1583 | 5.2239 |
87
- | 0.0 | 67.4603 | 8500 | 0.1587 | 5.2149 |
88
 
89
 
90
  ### Framework versions
 
24
  metrics:
25
  - name: Wer
26
  type: wer
27
+ value: 5.349683331087454
28
  ---
29
 
30
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
 
34
 
35
  This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the Physician Dictation GPT 4 Turbo dataset.
36
  It achieves the following results on the evaluation set:
37
+ - Loss: 0.1660
38
+ - Wer: 5.3497
39
 
40
  ## Model description
41
 
 
60
  - seed: 42
61
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
62
  - lr_scheduler_type: linear
 
63
  - training_steps: 8500
64
  - mixed_precision_training: Native AMP
65
 
 
67
 
68
  | Training Loss | Epoch | Step | Validation Loss | Wer |
69
  |:-------------:|:-------:|:----:|:---------------:|:------:|
70
+ | 0.0143 | 3.9683 | 500 | 0.1068 | 4.9948 |
71
+ | 0.0023 | 7.9365 | 1000 | 0.1287 | 5.0757 |
72
+ | 0.0026 | 11.9048 | 1500 | 0.1254 | 4.9409 |
73
+ | 0.0003 | 15.8730 | 2000 | 0.1298 | 4.7298 |
74
+ | 0.001 | 19.8413 | 2500 | 0.1312 | 5.0173 |
75
+ | 0.0001 | 23.8095 | 3000 | 0.1405 | 5.2374 |
76
+ | 0.0001 | 27.7778 | 3500 | 0.1454 | 4.9903 |
77
+ | 0.0001 | 31.7460 | 4000 | 0.1497 | 5.2104 |
78
+ | 0.0 | 35.7143 | 4500 | 0.1531 | 5.1835 |
79
+ | 0.0 | 39.6825 | 5000 | 0.1558 | 5.1431 |
80
+ | 0.0 | 43.6508 | 5500 | 0.1581 | 5.1296 |
81
+ | 0.0 | 47.6190 | 6000 | 0.1601 | 5.1700 |
82
+ | 0.0 | 51.5873 | 6500 | 0.1619 | 5.1925 |
83
+ | 0.0 | 55.5556 | 7000 | 0.1635 | 5.2329 |
84
+ | 0.0 | 59.5238 | 7500 | 0.1648 | 5.2733 |
85
+ | 0.0 | 63.4921 | 8000 | 0.1656 | 5.3362 |
86
+ | 0.0 | 67.4603 | 8500 | 0.1660 | 5.3497 |
87
 
88
 
89
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c084dfc4bd388f23b2d7a2250fc15de47f74eef2585795561a2fe54ff754dcde
3
  size 3055544304
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1062185a03c96618194db013f692aa70c2215a30eb3dcf305e9756c39b23503d
3
  size 3055544304