vadhri/distilhubert-finetuned-gtzan

music-genre-classifer-20-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: 1.1035
• Accuracy: 0.87

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: 4
• 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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0544	1.0	225	1.9608	0.47
1.2995	2.0	450	1.3852	0.51
0.8875	3.0	675	0.9288	0.71
0.4092	4.0	900	0.8114	0.76
0.5624	5.0	1125	0.8704	0.77
0.0609	6.0	1350	0.7951	0.82
0.1018	7.0	1575	0.7055	0.86
0.2941	8.0	1800	0.8832	0.83
0.0044	9.0	2025	0.9883	0.83
0.0025	10.0	2250	0.9306	0.88
0.0016	11.0	2475	0.9535	0.86
0.0012	12.0	2700	1.0921	0.85
0.001	13.0	2925	1.0428	0.86
0.0011	14.0	3150	1.2270	0.83
0.0008	15.0	3375	1.1831	0.84
0.0007	16.0	3600	1.2124	0.84
0.0007	17.0	3825	1.0806	0.86
0.2454	18.0	4050	1.1530	0.85
0.0006	19.0	4275	1.1078	0.86
0.0006	20.0	4500	1.1035	0.87

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2