Whisper tiny Chinese English
This model is a fine-tuned version of openai/whisper-tiny on the Chinese English dataset. It achieves the following results on the evaluation set:
- Loss: 0.4022
- Wer: 15.5997
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: 3.75e-05
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- 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
- training_steps: 1000
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2237 | 2.0408 | 100 | 0.3217 | 13.9793 |
| 0.0608 | 4.0816 | 200 | 0.3430 | 14.4559 |
| 0.0377 | 6.1224 | 300 | 0.3745 | 14.9007 |
| 0.0184 | 8.1633 | 400 | 0.3866 | 18.4432 |
| 0.0182 | 10.2041 | 500 | 0.3872 | 15.2820 |
| 0.0125 | 12.2449 | 600 | 0.3879 | 15.3932 |
| 0.0085 | 14.2857 | 700 | 0.4032 | 15.7427 |
| 0.0066 | 16.3265 | 800 | 0.4051 | 19.3010 |
| 0.0058 | 18.3673 | 900 | 0.4010 | 19.4281 |
| 0.0049 | 20.4082 | 1000 | 0.4022 | 15.5997 |
Framework versions
- Transformers 5.23.0.dev0
- Pytorch 2.11.0+cu130
- Datasets 5.1.0
- Tokenizers 0.23.3