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: 0.4772
• 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
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1579	1.0	113	2.0599	0.49
1.414	2.0	226	1.4245	0.66
1.1784	3.0	339	1.1034	0.65
0.7209	4.0	452	0.7965	0.78
0.5967	5.0	565	0.7356	0.78
0.3644	6.0	678	0.5675	0.82
0.2092	7.0	791	0.5001	0.86
0.0647	8.0	904	0.6535	0.83
0.1354	9.0	1017	0.4772	0.88
0.0525	10.0	1130	0.6393	0.86
0.0127	11.0	1243	0.6587	0.87
0.0059	12.0	1356	0.6669	0.84
0.0052	13.0	1469	0.6354	0.86
0.0044	14.0	1582	0.7918	0.82
0.0035	15.0	1695	0.7577	0.83
0.0035	16.0	1808	0.7344	0.82
0.0033	17.0	1921	0.7079	0.84
0.0031	18.0	2034	0.7367	0.84
0.0032	19.0	2147	0.7150	0.84
0.0034	20.0	2260	0.7144	0.84

Framework versions

• Transformers 4.52.4
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1