metadata
language:
- fr
license: apache-2.0
base_model: openai/whisper-small
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: openai/whisper-small
results: []
openai/whisper-small
This model is a fine-tuned version of openai/whisper-small on the pphuc25/FrenchMed dataset. It achieves the following results on the evaluation set:
- Loss: 1.4773
- Wer: 36.7302
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 |
|---|---|---|---|---|
| 1.0463 | 1.0 | 215 | 1.0179 | 55.0587 |
| 0.6082 | 2.0 | 430 | 1.0745 | 57.8446 |
| 0.3176 | 3.0 | 645 | 1.1829 | 41.1290 |
| 0.1609 | 4.0 | 860 | 1.2342 | 41.7155 |
| 0.1106 | 5.0 | 1075 | 1.2716 | 39.7361 |
| 0.0895 | 6.0 | 1290 | 1.3019 | 40.3226 |
| 0.0761 | 7.0 | 1505 | 1.3814 | 42.0088 |
| 0.056 | 8.0 | 1720 | 1.4171 | 42.3754 |
| 0.0379 | 9.0 | 1935 | 1.4578 | 40.6891 |
| 0.0278 | 10.0 | 2150 | 1.4600 | 40.2493 |
| 0.0235 | 11.0 | 2365 | 1.4685 | 37.9765 |
| 0.0134 | 12.0 | 2580 | 1.4823 | 40.1760 |
| 0.0129 | 13.0 | 2795 | 1.4950 | 39.4428 |
| 0.008 | 14.0 | 3010 | 1.4921 | 39.0762 |
| 0.0038 | 15.0 | 3225 | 1.4791 | 36.0704 |
| 0.0042 | 16.0 | 3440 | 1.4517 | 39.4428 |
| 0.0025 | 17.0 | 3655 | 1.4607 | 37.9032 |
| 0.0006 | 18.0 | 3870 | 1.4702 | 36.2903 |
| 0.0002 | 19.0 | 4085 | 1.4753 | 36.6569 |
| 0.0006 | 20.0 | 4300 | 1.4773 | 36.7302 |
Framework versions
- Transformers 4.41.1
- Pytorch 2.3.0
- Datasets 2.19.1
- Tokenizers 0.19.1