Whisper Small CKB - Pro (Best Step)
This model is a fine-tuned version of openai/whisper-small on the Google FLEURS dataset. It achieves the following results on the evaluation set:
- Loss: 0.3767
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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 |
|---|---|---|---|
| 0.2672 | 2.5907 | 500 | 0.3034 |
| 0.0416 | 5.1813 | 1000 | 0.2682 |
| 0.0152 | 7.7720 | 1500 | 0.2846 |
| 0.0032 | 10.3627 | 2000 | 0.3132 |
| 0.0013 | 12.9534 | 2500 | 0.3376 |
| 0.0006 | 15.5440 | 3000 | 0.3526 |
| 0.0003 | 18.1347 | 3500 | 0.3617 |
| 0.0002 | 20.7254 | 4000 | 0.3688 |
| 0.0001 | 23.3161 | 4500 | 0.3746 |
| 0.0002 | 25.9067 | 5000 | 0.3767 |
Framework versions
- Transformers 4.57.3
- Pytorch 2.9.0+cu126
- Datasets 2.19.0
- Tokenizers 0.22.1