Hugging Face's logo

linshoufan
/
linshoufan-whisper-small-nan-tw

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
Min Nan Chinese
whisper
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
linshoufan commited on
Commit
a44ebc2
•
1 Parent(s): ac1fc68

End of training

Browse files
Files changed (2) hide show
  1. README.md +10 -11
  2. runs/Mar11_07-22-26_112faff8b56e/events.out.tfevents.1710141749.112faff8b56e.26.0 +2 -2
README.md CHANGED
@@ -4,24 +4,24 @@ language:
4
  license: apache-2.0
5
  base_model: openai/whisper-small
6
  tags:
7
- - hf-asr-leaderboard
8
  - generated_from_trainer
9
  datasets:
 
10
  - mozilla-foundation/common_voice_16_1
11
  model-index:
12
- - name: Whisper Small nan-tw - Taiwanese
13
  results: []
14
  ---
15
 
16
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
  should probably proofread and complete it, then remove this comment. -->
18
 
19
- # Whisper Small nan-tw - Taiwanese
20
 
21
- This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 16.1 dataset.
22
  It achieves the following results on the evaluation set:
23
- - Loss: 0.7966
24
- - Cer: 44.6503
25
 
26
  ## Model description
27
 
@@ -47,17 +47,16 @@ The following hyperparameters were used during training:
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
  - lr_scheduler_warmup_steps: 500
50
- - training_steps: 4000
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
  | Training Loss | Epoch | Step | Validation Loss | Cer |
56
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
57
- | 0.329 | 2.99 | 1000 | 0.7453 | 48.6010 |
58
- | 0.0121 | 5.99 | 2000 | 0.7780 | 46.0104 |
59
- | 0.0013 | 8.98 | 3000 | 0.7884 | 44.5855 |
60
- | 0.0006 | 11.98 | 4000 | 0.7966 | 44.6503 |
61
 
62
 
63
  ### Framework versions
 
4
  license: apache-2.0
5
  base_model: openai/whisper-small
6
  tags:
 
7
  - generated_from_trainer
8
  datasets:
9
+ - mozilla-foundation/common_voice_15_0
10
  - mozilla-foundation/common_voice_16_1
11
  model-index:
12
+ - name: Whisper Small Taiwanese
13
  results: []
14
  ---
15
 
16
  <!-- This model card has been generated automatically according to the information the Trainer had access to. You
17
  should probably proofread and complete it, then remove this comment. -->
18
 
19
+ # Whisper Small Taiwanese
20
 
21
+ This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 15.0 and 16.1 dataset.
22
  It achieves the following results on the evaluation set:
23
+ - Loss: 0.4173
24
+ - Cer: 32.1243
25
 
26
  ## Model description
27
 
 
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
  - lr_scheduler_warmup_steps: 500
50
+ - num_epochs: 1
51
  - mixed_precision_training: Native AMP
52
 
53
  ### Training results
54
 
55
  | Training Loss | Epoch | Step | Validation Loss | Cer |
56
  |:-------------:|:-----:|:----:|:---------------:|:-------:|
57
+ | 0.5843 | 0.32 | 1000 | 0.6384 | 44.0259 |
58
+ | 0.3831 | 0.64 | 2000 | 0.4756 | 35.1447 |
59
+ | 0.3393 | 0.96 | 3000 | 0.4173 | 32.1243 |
 
60
 
61
 
62
  ### Framework versions
runs/Mar11_07-22-26_112faff8b56e/events.out.tfevents.1710141749.112faff8b56e.26.0 CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:f0a1c6f3ddcfc51417ae0b00b034b4fa3c720e5996511915ff52a8391d2cd54e
3
- size 31694
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9c6fdb5cdc2ceddeca7b0ede2c0d08358839b8f91210b4ef55e76ccdfc26f706
3
+ size 32892