Whisper Small Irish - Ino van de Wouw
This model is a fine-tuned version of openai/whisper-small on the Wikimedia-Speech-Irish dataset. It achieves the following results on the evaluation set:
- Loss: 0.2792
- Wer: 21.5510
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 2500
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.6959 | 0.6623 | 500 | 0.6916 | 45.1903 |
| 0.3564 | 1.3245 | 1000 | 0.4082 | 27.8983 |
| 0.2594 | 1.9868 | 1500 | 0.3233 | 23.9319 |
| 0.1527 | 2.6490 | 2000 | 0.2916 | 22.0176 |
| 0.1062 | 3.3113 | 2500 | 0.2792 | 21.5510 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.6.0+cu126
- Datasets 2.18.0
- Tokenizers 0.21.0