distilhubert-finetuned-gtzan
This model is a fine-tuned version of ntu-spml/distilhubert on the GTZAN dataset. It achieves the following results on the evaluation set:
- Loss: 1.7061
- Accuracy: 0.86
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: 0.0002
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- 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: cosine
- lr_scheduler_warmup_steps: 10
- num_epochs: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.9488 | 1.0 | 57 | 2.6518 | 0.58 |
| 2.0315 | 2.0 | 114 | 2.0176 | 0.7 |
| 1.2694 | 3.0 | 171 | 1.6823 | 0.75 |
| 1.1568 | 4.0 | 228 | 1.6164 | 0.81 |
| 1.3392 | 5.0 | 285 | 2.0431 | 0.68 |
| 0.4307 | 6.0 | 342 | 1.3216 | 0.82 |
| 0.5567 | 7.0 | 399 | 2.3048 | 0.74 |
| 0.2289 | 8.0 | 456 | 1.5631 | 0.85 |
| 0.0155 | 9.0 | 513 | 1.7105 | 0.83 |
| 0.0063 | 10.0 | 570 | 2.0512 | 0.83 |
| 0.0051 | 11.0 | 627 | 1.6647 | 0.86 |
| 0.0037 | 12.0 | 684 | 1.7787 | 0.85 |
| 0.0032 | 13.0 | 741 | 1.7046 | 0.86 |
| 0.0088 | 14.0 | 798 | 1.7008 | 0.86 |
| 0.0085 | 15.0 | 855 | 1.7061 | 0.86 |
Framework versions
- Transformers 5.23.0.dev0
- Pytorch 2.11.0+cu128
- Datasets 3.6.0
- Tokenizers 0.23.1