whisper-medium-chinese-tw-minnan-soap-demucs-params-by-gpt-freeze-encoder
This model is a fine-tuned version of openai/whisper-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.3884
- Wer: 200.8187
- Cer: 122.2363
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: 5e-06
- train_batch_size: 16
- eval_batch_size: 16
- 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: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.3608 | 4.6512 | 1000 | 1.8699 | 151.4035 | 103.4759 |
| 0.0123 | 9.3023 | 2000 | 2.2378 | 166.5107 | 101.4489 |
| 0.0033 | 13.9535 | 3000 | 2.3884 | 200.8187 | 122.2363 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.5.0
- Tokenizers 0.21.1