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: 0.9233
- 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: 5e-05
- 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: 100
- num_epochs: 50
- mixed_precision_training: Native AMP
- label_smoothing_factor: 0.1
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 2.2511 | 1.0 | 57 | 2.1960 | 0.35 |
| 1.8151 | 2.0 | 114 | 1.7180 | 0.56 |
| 1.4443 | 3.0 | 171 | 1.3887 | 0.67 |
| 1.1114 | 4.0 | 228 | 1.2480 | 0.71 |
| 1.0100 | 5.0 | 285 | 1.0616 | 0.76 |
| 0.8755 | 6.0 | 342 | 1.0482 | 0.8 |
| 0.7948 | 7.0 | 399 | 0.9815 | 0.81 |
| 0.6078 | 8.0 | 456 | 1.0874 | 0.78 |
| 0.6020 | 9.0 | 513 | 0.9195 | 0.83 |
| 0.5687 | 10.0 | 570 | 0.9876 | 0.79 |
| 0.5556 | 11.0 | 627 | 1.0149 | 0.8 |
| 0.5114 | 12.0 | 684 | 1.0709 | 0.79 |
| 0.5036 | 13.0 | 741 | 0.9233 | 0.86 |
| 0.5024 | 14.0 | 798 | 1.0685 | 0.81 |
| 0.5019 | 15.0 | 855 | 0.9641 | 0.85 |
| 0.5014 | 16.0 | 912 | 0.9102 | 0.85 |
| 0.5022 | 17.0 | 969 | 0.8981 | 0.85 |
| 0.5045 | 18.0 | 1026 | 0.9275 | 0.83 |
Framework versions
- Transformers 5.16.1
- Pytorch 2.14.1+cu130
- Datasets 5.1.0
- Tokenizers 0.23.3