Whisper Small - Dialek Badung
This model is a fine-tuned version of openai/whisper-small on the Dataset Bahasa Bali Dialek Badung dataset. It achieves the following results on the evaluation set:
- Loss: 0.3195
- Wer: 37.2706
- Cer: 18.4622
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: 200
- training_steps: 1000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.4088 | 2.6358 | 200 | 0.3205 | 55.8486 | 32.9359 |
| 0.0413 | 5.2649 | 400 | 0.2954 | 60.8945 | 32.4630 |
| 0.0047 | 7.9007 | 600 | 0.3212 | 37.8440 | 15.7278 |
| 0.0010 | 10.5298 | 800 | 0.3161 | 37.9587 | 18.8939 |
| 0.0008 | 13.1589 | 1000 | 0.3195 | 37.2706 | 18.4622 |
Framework versions
- Transformers 5.18.0
- Pytorch 2.11.0+cu130
- Datasets 4.8.5
- Tokenizers 0.23.2