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: 0.7162
- 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: 0.0001
- train_batch_size: 8
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-05
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 8
- label_smoothing_factor: 0.05
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.5923 | 1.0 | 113 | 1.7310 | 0.44 |
| 1.2071 | 2.0 | 226 | 1.2546 | 0.62 |
| 1.0673 | 3.0 | 339 | 0.9320 | 0.76 |
| 0.8149 | 4.0 | 452 | 0.8768 | 0.81 |
| 0.4999 | 5.0 | 565 | 0.7154 | 0.86 |
| 0.3562 | 6.0 | 678 | 0.6631 | 0.89 |
| 0.3852 | 7.0 | 791 | 0.7136 | 0.87 |
| 0.4476 | 8.0 | 904 | 0.7162 | 0.88 |
Framework versions
- Transformers 4.30.2
- Pytorch 2.0.1+cu118
- Datasets 2.13.1
- Tokenizers 0.13.3