Whisper-Small-kh
This model is a fine-tuned version of openai/whisper-small on the Fleur dataset. It achieves the following results on the evaluation set:
- Loss: 0.2773
- Wer Ortho: 40.6131
- Wer: 19.5100
- Cer: 10.6710
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: 64
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 128
- 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
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer | Cer |
|---|---|---|---|---|---|---|
| No log | 1.0 | 19 | 0.2679 | 40.6512 | 19.5904 | 10.7011 |
| 0.0006 | 2.0 | 38 | 0.2717 | 40.4037 | 19.5984 | 10.7388 |
| 0.0005 | 2.8649 | 54 | 0.2773 | 40.6131 | 19.5100 | 10.6710 |
Framework versions
- Transformers 4.56.2
- Pytorch 2.8.0+cu126
- Datasets 2.14.7
- Tokenizers 0.22.1