Whisper Small Isizulu-Asr-1.1-Speed
This model is a fine-tuned version of openai/whisper-small on the ISIZULU-ASR-1.1-SPEED dataset. It achieves the following results on the evaluation set:
- Loss: 1.3263
- Wer: 53.6
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: 32
- eval_batch_size: 16
- seed: 42
- optimizer: Use 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: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0004 | 55.5556 | 1000 | 1.2327 | 53.8286 |
| 0.0001 | 111.1111 | 2000 | 1.2957 | 53.8286 |
| 0.0001 | 166.6667 | 3000 | 1.3263 | 53.6 |
| 0.0001 | 222.2222 | 4000 | 1.3378 | 54.1714 |
Framework versions
- Transformers 4.57.0
- Pytorch 2.8.0+cu128
- Datasets 4.2.0
- Tokenizers 0.22.1