whisper-small-sna-candace
This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4515
- Wer: 0.2987
- Cer: 0.0758
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: 0.0001
- train_batch_size: 16
- eval_batch_size: 64
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- 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
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 0.6428 | 1.1710 | 500 | 0.4338 | 0.3262 | 0.0907 |
| 0.4696 | 2.3419 | 1000 | 0.3893 | 0.3164 | 0.1014 |
| 0.3525 | 3.5129 | 1500 | 0.4012 | 0.3356 | 0.1177 |
| 0.2624 | 4.6838 | 2000 | 0.4396 | 0.3271 | 0.0956 |
| 0.1856 | 5.8548 | 2500 | 0.4515 | 0.2987 | 0.0758 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2