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: 0.8080
• 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: 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
1.7263	1.0	113	1.6331	0.59
1.1167	2.0	226	1.2046	0.65
0.774	3.0	339	0.8365	0.8
0.6507	4.0	452	0.7015	0.81
0.5046	5.0	565	0.6722	0.81
0.2632	6.0	678	0.6743	0.82
0.203	7.0	791	0.7351	0.84
0.0902	8.0	904	0.5898	0.86
0.0215	9.0	1017	0.6213	0.87
0.0097	10.0	1130	0.6948	0.86
0.1171	11.0	1243	0.6228	0.87
0.0054	12.0	1356	0.7101	0.86
0.0035	13.0	1469	0.7626	0.87
0.0028	14.0	1582	0.7659	0.86
0.0027	15.0	1695	0.6993	0.87
0.0023	16.0	1808	0.7345	0.87
0.0023	17.0	1921	0.8363	0.86
0.0018	18.0	2034	0.7779	0.88
0.0018	19.0	2147	0.8275	0.87
0.0018	20.0	2260	0.8080	0.88

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.0+cu121
• Datasets 2.16.0
• Tokenizers 0.15.0