whisper-small-finetuned-gtzan
This model is a fine-tuned version of openai/whisper-small on the GTZAN dataset. It achieves the following results on the evaluation set:
- Loss: 0.3287
- Accuracy: 0.94
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: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- 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_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 7.3893 | 1.0 | 29 | 1.3379 | 0.63 |
| 2.9131 | 2.0 | 58 | 0.5808 | 0.85 |
| 1.217 | 3.0 | 87 | 0.5240 | 0.79 |
| 1.2575 | 4.0 | 116 | 0.4218 | 0.89 |
| 0.3431 | 5.0 | 145 | 0.4044 | 0.89 |
| 0.1446 | 6.0 | 174 | 0.4010 | 0.9 |
| 0.1352 | 7.0 | 203 | 0.5442 | 0.89 |
| 0.0107 | 8.0 | 232 | 0.2855 | 0.96 |
| 0.0073 | 9.0 | 261 | 0.3288 | 0.93 |
| 0.007 | 9.6726 | 280 | 0.3287 | 0.94 |
Framework versions
- Transformers 4.47.1
- Pytorch 2.5.1+cu121
- Datasets 3.2.0
- Tokenizers 0.21.0