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: 0.9399
• Accuracy: 0.83

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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1679	1.0	113	2.0910	0.38
1.4665	2.0	226	1.4798	0.53
1.2128	3.0	339	1.1715	0.64
0.7499	4.0	452	0.9591	0.68
0.6869	5.0	565	0.8078	0.76
0.3399	6.0	678	0.7513	0.81
0.3071	7.0	791	0.6606	0.84
0.0791	8.0	904	0.6416	0.84
0.1047	9.0	1017	0.7613	0.82
0.0784	10.0	1130	0.8558	0.82
0.0097	11.0	1243	0.9087	0.82
0.0071	12.0	1356	0.9155	0.83
0.0052	13.0	1469	0.9210	0.85
0.0044	14.0	1582	0.9543	0.84
0.0035	15.0	1695	0.9726	0.85
0.0032	16.0	1808	0.9183	0.84
0.0029	17.0	1921	0.9181	0.83
0.0027	18.0	2034	0.9575	0.84
0.0027	19.0	2147	0.9427	0.83
0.0026	20.0	2260	0.9399	0.83

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.2+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2