IoanaLivia/real_data_1_h_synth_standard_B_horoscope-whisper-small
This model is a fine-tuned version of openai/whisper-small on the IoanaLivia/real_data_1_h_synth_standard_B_horoscope dataset. It achieves the following results on the evaluation set:
- Loss: 0.2705
- Wer: 18.4459
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: 4
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- 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
- num_epochs: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0 | 0 | 0.4170 | 30.1014 |
| 0.3332 | 1.0 | 42 | 0.2841 | 19.8556 |
| 0.1356 | 2.0 | 84 | 0.2705 | 18.4459 |
| 0.0678 | 3.0 | 126 | 0.2814 | 18.9789 |
| 0.0331 | 4.0 | 168 | 0.3037 | 20.0791 |
| 0.0162 | 5.0 | 210 | 0.3249 | 19.4774 |
| 0.0098 | 6.0 | 252 | 0.3424 | 19.2711 |
| 0.0065 | 7.0 | 294 | 0.3458 | 18.8413 |
| 0.005 | 8.0 | 336 | 0.3524 | 19.0992 |
| 0.0039 | 9.0 | 378 | 0.3605 | 19.0992 |
| 0.0034 | 10.0 | 420 | 0.3654 | 18.8929 |
| 0.003 | 11.0 | 462 | 0.3680 | 18.9445 |
| 0.0028 | 12.0 | 504 | 0.3708 | 18.8929 |
| 0.0026 | 13.0 | 546 | 0.3729 | 19.0132 |
| 0.0025 | 14.0 | 588 | 0.3737 | 19.0304 |
| 0.0025 | 14.6506 | 615 | 0.3738 | 19.0132 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1