Whisper-medium-Multilingual
This model is a fine-tuned version of openai/whisper-medium on the combined_voice_dataset_en_ja dataset. It achieves the following results on the evaluation set:
- Loss: 0.2195
- Wer: 14.1862
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: 32
- 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: 500
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.4567 | 3.6364 | 1000 | 0.2337 | 28.9944 |
| 0.0403 | 7.2727 | 2000 | 0.2126 | 18.1510 |
| 0.0137 | 10.9091 | 3000 | 0.2142 | 17.1510 |
| 0.0013 | 14.5455 | 4000 | 0.2149 | 15.5248 |
| 0.0006 | 18.1818 | 5000 | 0.2195 | 14.1862 |
Framework versions
- Transformers 5.15.1
- Pytorch 2.13.0+cu130
- Datasets 2.21.0
- Tokenizers 0.22.2