Hugging Face's logo

AlinaKozyreva
/
whisper-small-ru

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Russian
whisper
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Metrics Community
AlinaKozyreva commited on
Commit
fcd8937
•
1 Parent(s): ec3e6ab

End of training

Browse files
Files changed (2) hide show
  1. README.md +15 -8
  2. model.safetensors +1 -1
README.md CHANGED
@@ -10,7 +10,7 @@ datasets:
10
  metrics:
11
  - wer
12
  model-index:
13
- - name: Whisper Small Ru - Model_ru_1
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
@@ -24,19 +24,19 @@ model-index:
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 14.953271028037381
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 Small Ru - Model_ru_1
34
 
35
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 13 dataset.
36
  It achieves the following results on the evaluation set:
37
- - Loss: 0.2366
38
- - Wer Ortho: 19.9130
39
- - Wer: 14.9533
40
 
41
  ## Model description
42
 
@@ -62,14 +62,21 @@ The following hyperparameters were used during training:
62
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
63
  - lr_scheduler_type: constant_with_warmup
64
  - lr_scheduler_warmup_steps: 50
65
- - training_steps: 500
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
  | Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
71
  |:-------------:|:-----:|:----:|:---------------:|:---------:|:-------:|
72
- | 0.209 | 0.22 | 500 | 0.2366 | 19.9130 | 14.9533 |
 
 
 
 
 
 
 
73
 
74
 
75
  ### Framework versions
 
10
  metrics:
11
  - wer
12
  model-index:
13
+ - name: Whisper Small Ru - Model_ru_3
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
 
24
  metrics:
25
  - name: Wer
26
  type: wer
27
+ value: 13.30140186915888
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 Small Ru - Model_ru_3
34
 
35
  This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 13 dataset.
36
  It achieves the following results on the evaluation set:
37
+ - Loss: 0.2080
38
+ - Wer Ortho: 17.4462
39
+ - Wer: 13.3014
40
 
41
  ## Model description
42
 
 
62
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
63
  - lr_scheduler_type: constant_with_warmup
64
  - lr_scheduler_warmup_steps: 50
65
+ - training_steps: 4000
66
  - mixed_precision_training: Native AMP
67
 
68
  ### Training results
69
 
70
  | Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
71
  |:-------------:|:-----:|:----:|:---------------:|:---------:|:-------:|
72
+ | 0.2085 | 0.22 | 500 | 0.2366 | 19.9234 | 14.9498 |
73
+ | 0.1875 | 0.44 | 1000 | 0.2176 | 19.3079 | 14.5643 |
74
+ | 0.1688 | 0.66 | 1500 | 0.2095 | 18.3736 | 13.9287 |
75
+ | 0.1678 | 0.88 | 2000 | 0.2038 | 17.7325 | 13.4381 |
76
+ | 0.0853 | 1.1 | 2500 | 0.2036 | 17.0309 | 12.7488 |
77
+ | 0.0822 | 1.32 | 3000 | 0.2046 | 17.6894 | 13.2780 |
78
+ | 0.0775 | 1.54 | 3500 | 0.2051 | 16.9948 | 12.7126 |
79
+ | 0.0727 | 1.76 | 4000 | 0.2080 | 17.4462 | 13.3014 |
80
 
81
 
82
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e4e71fc18b450b4b711bc9e6683a5b86dc810354edb0abd5f7394807b6280bec
3
  size 966995080
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:46acacf9f7431093eccc8e2a9b8a99fc7fbecc866d9f13cb16465d3e1390628b
3
  size 966995080