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: 0.7565
• 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: 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
2.1839	1.0	113	2.0630	0.41
1.5052	2.0	226	1.4029	0.57
1.144	3.0	339	0.9807	0.77
0.9971	4.0	452	0.8701	0.75
0.6168	5.0	565	0.7094	0.76
0.4665	6.0	678	0.5940	0.83
0.58	7.0	791	0.4763	0.86
0.1009	8.0	904	0.4859	0.87
0.1817	9.0	1017	0.5313	0.88
0.0467	10.0	1130	0.6114	0.86
0.0201	11.0	1243	0.6677	0.85
0.1188	12.0	1356	0.6934	0.87
0.0055	13.0	1469	0.7070	0.89
0.0046	14.0	1582	0.7601	0.87
0.0043	15.0	1695	0.7584	0.87
0.0033	16.0	1808	0.7588	0.86
0.0696	17.0	1921	0.7495	0.88
0.0028	18.0	2034	0.7535	0.87
0.0027	19.0	2147	0.7571	0.87
0.0028	20.0	2260	0.7565	0.87

Framework versions

• Transformers 4.37.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0