Quick-trained-trial Whisper-Small Yue model #JL
This model is a fine-tuned version of openai/whisper-small on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.3101
- Cer: 12.7086
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: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Cer |
|---|---|---|---|---|
| 0.0646 | 2.7778 | 1000 | 0.2421 | 14.9583 |
| 0.008 | 5.5556 | 2000 | 0.2735 | 18.6681 |
| 0.0007 | 8.3333 | 3000 | 0.3033 | 15.4648 |
| 0.0005 | 11.1111 | 4000 | 0.3101 | 12.7086 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.7.0+cu126
- Datasets 3.5.1
- Tokenizers 0.21.1