whisper-large-v3-persian
This model is a fine-tuned version of openai/whisper-large-v3 on the ASR_fa_v1 dataset. It achieves the following results on the evaluation set:
- Loss: 0.0858
- Wer: 8.7300
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: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.1501 | 0.5970 | 1000 | 0.1537 | 17.1059 |
| 0.081 | 1.1940 | 2000 | 0.1156 | 12.6248 |
| 0.0766 | 1.7910 | 3000 | 0.0965 | 11.1969 |
| 0.0313 | 2.3881 | 4000 | 0.0877 | 9.3975 |
| 0.0263 | 2.9851 | 5000 | 0.0858 | 8.7300 |
Framework versions
- Transformers 4.52.4
- Pytorch 2.7.0+cu126
- Datasets 3.6.0
- Tokenizers 0.21.1