whisper-small-pashto
This model is a fine-tuned version of YuvrajGujari/whisper-small-pashto on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5657
- Wer: 43.5138
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: 5e-06
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 50
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.5598 | 0.6614 | 250 | 0.5347 | 44.3038 |
| 0.4469 | 1.3228 | 500 | 0.5397 | 43.0094 |
| 0.4867 | 1.9841 | 750 | 0.5216 | 42.8762 |
| 0.3296 | 2.6455 | 1000 | 0.5379 | 43.2093 |
| 0.2653 | 3.3069 | 1250 | 0.5462 | 43.2759 |
| 0.2619 | 3.9683 | 1500 | 0.5496 | 43.2188 |
| 0.2001 | 4.6296 | 1750 | 0.5618 | 43.6185 |
| 0.1736 | 5.2910 | 2000 | 0.5657 | 43.5138 |
Framework versions
- Transformers 5.14.1
- Pytorch 2.8.0+cu128
- Datasets 2.19.2
- Tokenizers 0.22.2