distilhubert-finetuned-gtzan-v2
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: 0.4959
- Accuracy: 0.87
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 |
|---|---|---|---|---|
| 1.7507 | 1.0 | 113 | 1.8033 | 0.42 |
| 1.2592 | 2.0 | 226 | 1.1710 | 0.71 |
| 1.039 | 3.0 | 339 | 0.9022 | 0.73 |
| 0.6122 | 4.0 | 452 | 0.6954 | 0.82 |
| 0.4654 | 5.0 | 565 | 0.6944 | 0.84 |
| 0.2895 | 6.0 | 678 | 0.5393 | 0.85 |
| 0.2114 | 7.0 | 791 | 0.5197 | 0.86 |
| 0.1997 | 8.0 | 904 | 0.5195 | 0.85 |
| 0.1282 | 9.0 | 1017 | 0.4883 | 0.87 |
| 0.3031 | 10.0 | 1130 | 0.4959 | 0.87 |
Framework versions
- Transformers 4.29.2
- Pytorch 2.0.1+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3