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: 0.76
• 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: 4 - 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: 15 - .train_test_split(seed=2024, shuffle=True, test_size=0.1)

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
1.9415	1.0	113	0.55	1.8500
1.3078	2.0	226	0.58	1.3794
1.1238	3.0	339	0.65	1.0919
0.788	4.0	452	0.68	1.0212
0.5932	5.0	565	0.69	0.8691
0.4042	6.0	678	0.71	0.8527
0.3421	7.0	791	0.75	0.7737
0.223	8.0	904	0.75	0.8463
0.1162	9.0	1017	0.77	0.7808
0.0863	10.0	1130	0.75	0.7487
0.1357	11.0	1243	0.8839	0.76
0.0632	12.0	1356	0.7509	0.76
0.0342	13.0	1469	0.8219	0.77
0.0277	14.0	1582	0.7691	0.8
0.0307	15.0	1695	0.7854	0.77

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.2
• Datasets 2.16.1
• Tokenizers 0.13.2