Hugging Face's logo

vittai23
/
whisper-small-marathiV1

Automatic Speech Recognition
Transformers
PyTorch
TensorBoard
Marathi
whisper
hf-asr-leaderboard
generated_from_trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
vittai23 commited on
Commit
6d5dbe5
•
1 Parent(s): 4a5f81e

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +13 -14
README.md CHANGED
@@ -10,7 +10,7 @@ datasets:
10
  metrics:
11
  - wer
12
  model-index:
13
- - name: whisper_marathi_V2
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
@@ -24,18 +24,18 @@ model-index:
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 545.1292631036039
28
  ---
29
 
30
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
31
  should probably proofread and complete it, then remove this comment. -->
32
 
33
- # whisper_marathi_V2
34
 
35
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
36
  It achieves the following results on the evaluation set:
37
- - Loss: 0.4603
38
- - Wer: 545.1293
39
 
40
  ## Model description
41
 
@@ -60,23 +60,22 @@ 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: 4000
65
  - mixed_precision_training: Native AMP
66
 
67
  ### Training results
68
 
69
- | Training Loss | Epoch | Step | Validation Loss | Wer |
70
- |:-------------:|:-----:|:----:|:---------------:|:--------:|
71
- | 0.0673 | 4.07 | 1000 | 0.2908 | 100.4062 |
72
- | 0.0045 | 8.13 | 2000 | 0.3941 | 217.4973 |
73
- | 0.0003 | 12.2 | 3000 | 0.4377 | 474.5600 |
74
- | 0.0002 | 16.26 | 4000 | 0.4603 | 545.1293 |
75
 
76
 
77
  ### Framework versions
78
 
79
  - Transformers 4.27.0.dev0
80
  - Pytorch 1.13.1+cu116
81
- - Datasets 2.10.0
82
  - Tokenizers 0.13.2
 
10
  metrics:
11
  - wer
12
  model-index:
13
+ - name: whisper_marathi_test
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
 
24
  metrics:
25
  - name: Wer
26
  type: wer
27
+ value: 53.72961124363355
28
  ---
29
 
30
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
31
  should probably proofread and complete it, then remove this comment. -->
32
 
33
+ # whisper_marathi_test
34
 
35
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset.
36
  It achieves the following results on the evaluation set:
37
+ - Loss: 0.3127
38
+ - Wer: 53.7296
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
+ - lr_scheduler_warmup_steps: 10
64
+ - training_steps: 300
65
  - mixed_precision_training: Native AMP
66
 
67
  ### Training results
68
 
69
+ | Training Loss | Epoch | Step | Validation Loss | Wer |
70
+ |:-------------:|:-----:|:----:|:---------------:|:-------:|
71
+ | 0.4508 | 0.41 | 100 | 0.4453 | 60.3572 |
72
+ | 0.3239 | 0.81 | 200 | 0.3297 | 54.9159 |
73
+ | 0.24 | 1.22 | 300 | 0.3127 | 53.7296 |
 
74
 
75
 
76
  ### Framework versions
77
 
78
  - Transformers 4.27.0.dev0
79
  - Pytorch 1.13.1+cu116
80
+ - Datasets 2.10.1
81
  - Tokenizers 0.13.2