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: 0.6376
- Accuracy: 0.85
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: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- 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 |
|---|---|---|---|---|
| 1.7703 | 1.0 | 225 | 1.6440 | 0.45 |
| 0.9968 | 2.0 | 450 | 1.1709 | 0.6 |
| 0.3874 | 3.0 | 675 | 0.7769 | 0.77 |
| 0.8894 | 4.0 | 900 | 0.5280 | 0.84 |
| 0.1964 | 5.0 | 1125 | 0.6280 | 0.84 |
| 0.2273 | 6.0 | 1350 | 0.6823 | 0.82 |
| 0.0686 | 7.0 | 1575 | 0.6527 | 0.85 |
| 0.1212 | 8.0 | 1800 | 0.5111 | 0.86 |
| 0.014 | 9.0 | 2025 | 0.5715 | 0.86 |
| 0.012 | 10.0 | 2250 | 0.6376 | 0.85 |
Framework versions
- Transformers 4.39.2
- Pytorch 1.13.0+cu117
- Datasets 2.16.1
- Tokenizers 0.15.1