whisper-jangli-small
This model is a fine-tuned version of openai/whisper-small on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 0.9580
- Wer: 82.5225
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: 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
- num_epochs: 4
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.5977 | 1.8349 | 200 | 0.9239 | 85.5165 |
| 0.8364 | 3.6697 | 400 | 0.9561 | 82.5973 |
| 0.7979 | 4.0 | 436 | 0.9580 | 82.5225 |
Framework versions
- Transformers 5.20.1
- Pytorch 2.11.0+cu130
- Datasets 4.8.5
- Tokenizers 0.23.3