Whisper base uz - Jamshid Ahmadov
This model is a fine-tuned version of Whisper Base on an Common Voice dataset. It achieves the following results on the evaluation set:
- Loss: 0.1652
- Wer: 14.0135
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 1e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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
- lr_scheduler_warmup_steps: 500
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0346 | 0.5714 | 500 | 0.1719 | 14.7950 |
| 0.0348 | 1.1429 | 1000 | 0.1703 | 14.2490 |
| 0.0327 | 1.7143 | 1500 | 0.1672 | 14.1848 |
| 0.02 | 2.2857 | 2000 | 0.1652 | 14.0135 |
Framework versions
- Transformers 4.50.3
- Pytorch 2.5.1+cu121
- Datasets 3.5.0
- Tokenizers 0.21.0