whisper-tiny-finetuned-gtzan
This model is a fine-tuned version of openai/whisper-tiny on the GTZAN dataset. It achieves the following results on the evaluation set:
- Loss: 0.4198
- Accuracy: 0.88
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: 16
- eval_batch_size: 16
- seed: 42
- 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 |
|---|---|---|---|---|
| 1.7863 | 1.0 | 57 | 1.5165 | 0.64 |
| 0.9074 | 2.0 | 114 | 0.9433 | 0.67 |
| 0.5972 | 3.0 | 171 | 0.6179 | 0.8 |
| 0.3472 | 4.0 | 228 | 0.5855 | 0.78 |
| 0.2699 | 5.0 | 285 | 0.4670 | 0.84 |
| 0.1025 | 6.0 | 342 | 0.5236 | 0.81 |
| 0.0892 | 7.0 | 399 | 0.4453 | 0.85 |
| 0.0163 | 8.0 | 456 | 0.4244 | 0.91 |
| 0.0109 | 9.0 | 513 | 0.3771 | 0.9 |
| 0.01 | 10.0 | 570 | 0.4198 | 0.88 |
Framework versions
- Transformers 4.52.3
- Pytorch 2.6.0+cu124
- Datasets 2.14.4
- Tokenizers 0.21.1