openai/whisper-base
This model is a fine-tuned version of openai/whisper-base on the pphuc25/EngMed dataset. It achieves the following results on the evaluation set:
- Loss: 1.3225
- Wer: 27.9808
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 |
|---|---|---|---|---|
| 0.7862 | 1.0 | 3491 | 0.8811 | 43.2441 |
| 0.5443 | 2.0 | 6982 | 0.8834 | 38.7694 |
| 0.3319 | 3.0 | 10473 | 0.9247 | 34.6764 |
| 0.2412 | 4.0 | 13964 | 0.9765 | 34.9636 |
| 0.1809 | 5.0 | 17455 | 1.0223 | 31.3646 |
| 0.1112 | 6.0 | 20946 | 1.1013 | 33.5169 |
| 0.0786 | 7.0 | 24437 | 1.1217 | 34.1563 |
| 0.0666 | 8.0 | 27928 | 1.1698 | 36.5356 |
| 0.0487 | 9.0 | 31419 | 1.1934 | 33.9412 |
| 0.0395 | 10.0 | 34910 | 1.2259 | 31.4509 |
| 0.0229 | 11.0 | 38401 | 1.2614 | 32.0655 |
| 0.0181 | 12.0 | 41892 | 1.2823 | 30.5444 |
| 0.0118 | 13.0 | 45383 | 1.2890 | 30.2773 |
| 0.0069 | 14.0 | 48874 | 1.3081 | 30.3435 |
| 0.0073 | 15.0 | 52365 | 1.3085 | 30.4097 |
| 0.0017 | 16.0 | 55856 | 1.3099 | 29.4145 |
| 0.0062 | 17.0 | 59347 | 1.3229 | 29.5386 |
| 0.0013 | 18.0 | 62838 | 1.3162 | 28.3058 |
| 0.0001 | 19.0 | 66329 | 1.3197 | 28.0387 |
| 0.0001 | 20.0 | 69820 | 1.3225 | 27.9808 |
Framework versions
- Transformers 4.41.1
- Pytorch 2.3.0
- Datasets 2.19.1
- Tokenizers 0.19.1
- Downloads last month
- 11