xezpeleta commited on
Commit
b43bcdc
1 Parent(s): efeb365

update model card README.md

Browse files
Files changed (3) hide show
  1. .gitattributes +1 -0
  2. .run.sh.swp +3 -0
  3. README.md +28 -22
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ .run.sh.swp filter=lfs diff=lfs merge=lfs -text
.run.sh.swp ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:31c505fa1992c80920e5a50efa1d6e1dd9f6c9f7ae30cdd64d203eb6ce9ac573
3
+ size 12288
README.md CHANGED
@@ -1,41 +1,38 @@
1
  ---
2
- language:
3
- - eu
4
  license: apache-2.0
5
  tags:
6
- - whisper-event
7
  - generated_from_trainer
8
  datasets:
9
- - mozilla-foundation/common_voice_13_0
10
  metrics:
11
  - wer
12
  model-index:
13
- - name: Whisper Small Basque
14
  results:
15
  - task:
16
  name: Automatic Speech Recognition
17
  type: automatic-speech-recognition
18
  dataset:
19
- name: mozilla-foundation/common_voice_13_0 eu
20
- type: mozilla-foundation/common_voice_13_0
21
  config: eu
22
  split: test
23
  args: eu
24
  metrics:
25
  - name: Wer
26
  type: wer
27
- value: 13.179958686054519
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 Basque
34
 
35
- This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the mozilla-foundation/common_voice_13_0 eu dataset.
36
  It achieves the following results on the evaluation set:
37
- - Loss: 0.2201
38
- - Wer: 13.1800
39
 
40
  ## Model description
41
 
@@ -43,7 +40,7 @@ More information needed
43
 
44
  ## Intended uses & limitations
45
 
46
- If you need to use this model with [whisper.cpp](https://github.com/ggerganov/whisper.cpp), you can download the ggml file: [ggml-medium-eu.bin](https://huggingface.co/xezpeleta/whisper-medium-eu/blob/main/ggml-medium.eu.bin)
47
 
48
  ## Training and evaluation data
49
 
@@ -54,27 +51,36 @@ More information needed
54
  ### Training hyperparameters
55
 
56
  The following hyperparameters were used during training:
57
- - learning_rate: 1e-05
58
  - train_batch_size: 4
59
  - eval_batch_size: 8
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: 7000
65
  - mixed_precision_training: Native AMP
66
 
67
  ### Training results
68
 
69
  | Training Loss | Epoch | Step | Validation Loss | Wer |
70
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
71
- | 0.4203 | 0.14 | 1000 | 0.4128 | 28.2656 |
72
- | 0.2693 | 0.29 | 2000 | 0.3240 | 22.0523 |
73
- | 0.2228 | 0.43 | 3000 | 0.2737 | 18.1437 |
74
- | 0.1002 | 1.1 | 4000 | 0.2554 | 16.3534 |
75
- | 0.0863 | 1.24 | 5000 | 0.2351 | 14.7880 |
76
- | 0.0636 | 1.39 | 6000 | 0.2251 | 13.5971 |
77
- | 0.0271 | 2.06 | 7000 | 0.2201 | 13.1800 |
 
 
 
 
 
 
 
 
 
78
 
79
 
80
  ### Framework versions
 
1
  ---
 
 
2
  license: apache-2.0
3
  tags:
 
4
  - generated_from_trainer
5
  datasets:
6
+ - common_voice_13_0
7
  metrics:
8
  - wer
9
  model-index:
10
+ - name: openai/whisper-medium
11
  results:
12
  - task:
13
  name: Automatic Speech Recognition
14
  type: automatic-speech-recognition
15
  dataset:
16
+ name: common_voice_13_0
17
+ type: common_voice_13_0
18
  config: eu
19
  split: test
20
  args: eu
21
  metrics:
22
  - name: Wer
23
  type: wer
24
+ value: 14.119648426424725
25
  ---
26
 
27
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
28
  should probably proofread and complete it, then remove this comment. -->
29
 
30
+ # openai/whisper-medium
31
 
32
+ This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the common_voice_13_0 dataset.
33
  It achieves the following results on the evaluation set:
34
+ - Loss: 0.2376
35
+ - Wer: 14.1196
36
 
37
  ## Model description
38
 
 
40
 
41
  ## Intended uses & limitations
42
 
43
+ More information needed
44
 
45
  ## Training and evaluation data
46
 
 
51
  ### Training hyperparameters
52
 
53
  The following hyperparameters were used during training:
54
+ - learning_rate: 6e-06
55
  - train_batch_size: 4
56
  - eval_batch_size: 8
57
  - seed: 42
58
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
59
  - lr_scheduler_type: linear
60
  - lr_scheduler_warmup_steps: 500
61
+ - training_steps: 8000
62
  - mixed_precision_training: Native AMP
63
 
64
  ### Training results
65
 
66
  | Training Loss | Epoch | Step | Validation Loss | Wer |
67
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
68
+ | 0.443 | 0.06 | 500 | 0.5037 | 37.4296 |
69
+ | 0.4196 | 0.12 | 1000 | 0.4010 | 28.9137 |
70
+ | 0.2823 | 0.19 | 1500 | 0.3453 | 24.6851 |
71
+ | 0.2551 | 0.25 | 2000 | 0.3164 | 22.5789 |
72
+ | 0.206 | 0.31 | 2500 | 0.2902 | 19.7922 |
73
+ | 0.2327 | 0.38 | 3000 | 0.2707 | 18.9356 |
74
+ | 0.1416 | 1.03 | 3500 | 0.2566 | 17.6921 |
75
+ | 0.0998 | 1.09 | 4000 | 0.2551 | 16.8213 |
76
+ | 0.095 | 1.15 | 4500 | 0.2511 | 16.3899 |
77
+ | 0.0971 | 1.21 | 5000 | 0.2415 | 15.5393 |
78
+ | 0.0964 | 1.28 | 5500 | 0.2336 | 15.1707 |
79
+ | 0.072 | 1.34 | 6000 | 0.2353 | 14.7596 |
80
+ | 0.0658 | 1.4 | 6500 | 0.2340 | 14.6766 |
81
+ | 0.033 | 2.05 | 7000 | 0.2349 | 14.3768 |
82
+ | 0.0288 | 2.11 | 7500 | 0.2371 | 14.1865 |
83
+ | 0.0352 | 2.18 | 8000 | 0.2376 | 14.1196 |
84
 
85
 
86
  ### Framework versions