distilhubert-gtzan-baseline
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: 1.0109
- Accuracy: 0.84
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
- gradient_accumulation_steps: 2
- total_train_batch_size: 8
- 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: linear
- lr_scheduler_warmup_steps: 100
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.4207 | 1.0 | 113 | 0.8240 | 0.8 |
| 0.4825 | 2.0 | 226 | 0.8756 | 0.8 |
| 0.0528 | 3.0 | 339 | 0.7559 | 0.82 |
| 0.0715 | 4.0 | 452 | 1.1074 | 0.76 |
| 0.0111 | 5.0 | 565 | 0.9829 | 0.8 |
| 0.0074 | 6.0 | 678 | 1.1246 | 0.82 |
| 0.0020 | 7.0 | 791 | 1.0282 | 0.84 |
| 0.0015 | 8.0 | 904 | 0.9513 | 0.86 |
| 0.0014 | 9.0 | 1017 | 0.9888 | 0.85 |
| 0.0013 | 10.0 | 1130 | 1.0109 | 0.84 |
Framework versions
- Transformers 5.15.0
- Pytorch 2.11.0+cu128
- Datasets 2.21.0
- Tokenizers 0.22.2