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: 0.5726
- Accuracy: 0.82
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
- 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 |
|---|---|---|---|---|
| 1.9663 | 1.0 | 113 | 1.8383 | 0.5 |
| 1.3044 | 2.0 | 226 | 1.2317 | 0.72 |
| 0.9295 | 3.0 | 339 | 1.0049 | 0.72 |
| 0.6622 | 4.0 | 452 | 0.7796 | 0.81 |
| 0.4615 | 5.0 | 565 | 0.6917 | 0.8 |
| 0.3902 | 6.0 | 678 | 0.6734 | 0.83 |
| 0.3192 | 7.0 | 791 | 0.6424 | 0.84 |
| 0.1831 | 8.0 | 904 | 0.5786 | 0.84 |
| 0.1182 | 9.0 | 1017 | 0.5741 | 0.84 |
| 0.0965 | 10.0 | 1130 | 0.5726 | 0.82 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.0+cu121
- Datasets 2.16.0
- Tokenizers 0.15.0