openai/whisper-small
This model is a fine-tuned version of openai/whisper-small on the pphuc25/VietMed-split-8-2 dataset. It achieves the following results on the evaluation set:
- Loss: 0.9118
- Wer: 21.7024
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 20
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.6339 | 1.0 | 569 | 0.6214 | 27.8675 |
| 0.4322 | 2.0 | 1138 | 0.5860 | 25.1949 |
| 0.2846 | 3.0 | 1707 | 0.6092 | 24.0564 |
| 0.1624 | 4.0 | 2276 | 0.6697 | 24.2614 |
| 0.0915 | 5.0 | 2845 | 0.7291 | 25.9162 |
| 0.0642 | 6.0 | 3414 | 0.7756 | 24.7739 |
| 0.0382 | 7.0 | 3983 | 0.8095 | 24.1955 |
| 0.0355 | 8.0 | 4552 | 0.8362 | 24.2541 |
| 0.0314 | 9.0 | 5121 | 0.8388 | 23.9063 |
| 0.0174 | 10.0 | 5690 | 0.8719 | 22.9251 |
| 0.0107 | 11.0 | 6259 | 0.8740 | 23.0899 |
| 0.0109 | 12.0 | 6828 | 0.8811 | 22.8226 |
| 0.0036 | 13.0 | 7397 | 0.8770 | 22.5700 |
| 0.0037 | 14.0 | 7966 | 0.8888 | 22.4382 |
| 0.002 | 15.0 | 8535 | 0.8917 | 22.3467 |
| 0.0001 | 16.0 | 9104 | 0.9047 | 22.0099 |
| 0.0004 | 17.0 | 9673 | 0.8960 | 21.9660 |
| 0.0009 | 18.0 | 10242 | 0.9079 | 21.6914 |
| 0.0001 | 19.0 | 10811 | 0.9108 | 21.6658 |
| 0.0 | 20.0 | 11380 | 0.9118 | 21.7024 |
Framework versions
- Transformers 4.41.1
- Pytorch 2.3.0
- Datasets 2.19.1
- Tokenizers 0.19.1
- Downloads last month
- 0