Whisper Small Japanese - Your Name
This model is a fine-tuned version of openai/whisper-small on the Common Voice 8.0 Japanese dataset. It achieves the following results on the evaluation set:
- Loss: 0.3763
- Wer: 72.2467
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: 8000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0349 | 3.7453 | 1000 | 0.2838 | 71.8062 |
| 0.0031 | 7.4906 | 2000 | 0.3100 | 69.6035 |
| 0.0007 | 11.2360 | 3000 | 0.3358 | 70.9251 |
| 0.0003 | 14.9813 | 4000 | 0.3474 | 73.5683 |
| 0.0002 | 18.7266 | 5000 | 0.3555 | 73.1278 |
| 0.0002 | 22.4719 | 6000 | 0.3663 | 73.1278 |
| 0.0001 | 26.2172 | 7000 | 0.3732 | 72.2467 |
| 0.0001 | 29.9625 | 8000 | 0.3763 | 72.2467 |
Framework versions
- Transformers 5.1.0
- Pytorch 2.5.1+cu121
- Datasets 2.21.0
- Tokenizers 0.22.2
- Downloads last month
- 4
Safetensors
Model size
0.2B params
Tensor type
F32
·
Model tree for kdl02/whisper-small-ja
Base model
openai/whisper-smallDataset used to train kdl02/whisper-small-ja
Evaluation results
- Wer on Common Voice 8.0 Japaneseself-reported72.247