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: 0.6623
• Accuracy: 0.82

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: 3e-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
2.2457	1.0	113	2.1827	0.33
1.8385	2.0	226	1.6935	0.61
1.46	3.0	339	1.4282	0.63
1.1508	4.0	452	1.1055	0.7
0.9972	5.0	565	0.8945	0.74
0.7826	6.0	678	0.7784	0.77
0.6802	7.0	791	0.7184	0.8
0.4635	8.0	904	0.7725	0.76
0.3746	9.0	1017	0.5875	0.84
0.264	10.0	1130	0.7612	0.75
0.1995	11.0	1243	0.6099	0.81
0.135	12.0	1356	0.6306	0.81
0.0974	13.0	1469	0.5947	0.83
0.0563	14.0	1582	0.7485	0.8
0.0443	15.0	1695	0.6977	0.79
0.0565	16.0	1808	0.6331	0.83
0.0295	17.0	1921	0.6538	0.82
0.0178	18.0	2034	0.6977	0.82
0.0191	19.0	2147	0.6453	0.83
0.0147	20.0	2260	0.6623	0.82

Framework versions

• Transformers 4.41.2
• Pytorch 2.1.2
• Datasets 2.19.2
• Tokenizers 0.19.1