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: 0.7537
• 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: 16
• eval_batch_size: 16
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.9647	1.0	113	1.8614	0.52
1.3987	2.0	226	1.3098	0.61
0.8809	3.0	339	0.8631	0.76
0.7643	4.0	452	0.8114	0.77
0.5958	5.0	565	0.7013	0.81
0.4405	6.0	678	0.5860	0.84
0.2183	7.0	791	0.6114	0.82
0.1587	8.0	904	0.5141	0.85
0.0899	9.0	1017	0.4760	0.87
0.0575	10.0	1130	0.5759	0.86
0.0647	11.0	1243	0.6467	0.86
0.0061	12.0	1356	0.6372	0.88
0.0029	13.0	1469	0.6721	0.88
0.0018	14.0	1582	0.7565	0.89
0.0013	15.0	1695	0.7537	0.88

Framework versions

• Transformers 4.30.1
• Pytorch 2.0.1+cu117
• Datasets 2.14.0
• Tokenizers 0.13.3