Whisper Small Urdu V2 - Muhammad Abdullah
This model is a fine-tuned version of GogetaBlueMUI/whisper-small-ur on the Common Voice 13.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.7436
- Wer: 35.4231
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 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: 1500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.068 | 1.9305 | 500 | 0.6670 | 37.1751 |
| 0.0182 | 3.8610 | 1000 | 0.7094 | 35.9684 |
| 0.0032 | 5.7915 | 1500 | 0.7436 | 35.4231 |
Framework versions
- Transformers 4.47.0
- Pytorch 2.5.1+cu121
- Datasets 3.1.0
- Tokenizers 0.21.0