400-mixed-real-ro-RO-Standard-A
This model is a fine-tuned version of openai/whisper-small on the 400-mixed-real-ro-ro-standard-a dataset. It achieves the following results on the evaluation set:
- Loss: 0.3169
- Wer: 15.7571
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: 1e-05
- train_batch_size: 8
- eval_batch_size: 4
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0 | 0 | 0.6890 | 41.0325 |
| 0.2485 | 1.0 | 50 | 0.2640 | 17.4737 |
| 0.0577 | 2.0 | 100 | 0.2620 | 16.3976 |
| 0.0201 | 3.0 | 150 | 0.2813 | 16.6026 |
| 0.0105 | 4.0 | 200 | 0.2965 | 16.1158 |
| 0.0053 | 5.0 | 250 | 0.3080 | 15.8724 |
| 0.0029 | 6.0 | 300 | 0.3093 | 15.5265 |
| 0.0021 | 7.0 | 350 | 0.3124 | 15.7443 |
| 0.0019 | 8.0 | 400 | 0.3144 | 15.6546 |
| 0.0017 | 9.0 | 450 | 0.3164 | 15.7827 |
| 0.0016 | 10.0 | 500 | 0.3169 | 15.7571 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.5.1+cu121
- Datasets 3.4.1
- Tokenizers 0.21.0