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: 0.5321
• 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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1271	1.0	113	2.0529	0.47
1.4245	2.0	226	1.4173	0.6
1.1783	3.0	339	1.0567	0.71
0.7597	4.0	452	0.8387	0.75
0.6043	5.0	565	0.6876	0.81
0.4758	6.0	678	0.6897	0.79
0.4882	7.0	791	0.6507	0.79
0.2361	8.0	904	0.6232	0.84
0.209	9.0	1017	0.5800	0.82
0.0859	10.0	1130	0.5414	0.85
0.0639	11.0	1243	0.5321	0.88
0.0405	12.0	1356	0.8187	0.82
0.0481	13.0	1469	0.7086	0.85
0.0127	14.0	1582	0.7394	0.84
0.0071	15.0	1695	0.6890	0.86
0.0073	16.0	1808	0.7361	0.86
0.0062	17.0	1921	0.9311	0.8
0.0028	18.0	2034	0.7819	0.84
0.0024	19.0	2147	0.8263	0.86
0.0023	20.0	2260	0.8049	0.86

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0