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: 0.6333
• 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: 16
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2417	1.0	57	2.1896	0.42
1.8003	2.0	114	1.6369	0.52
1.3938	3.0	171	1.2560	0.72
1.2724	4.0	228	1.1942	0.68
0.9682	5.0	285	0.8864	0.8
0.7111	6.0	342	0.7542	0.82
0.6339	7.0	399	0.7712	0.81
0.4599	8.0	456	0.6080	0.84
0.3261	9.0	513	0.5998	0.84
0.2991	10.0	570	0.6767	0.79
0.1615	11.0	627	0.5817	0.87
0.0854	12.0	684	0.5859	0.83
0.0752	13.0	741	0.5681	0.85
0.0341	14.0	798	0.5916	0.88
0.0331	15.0	855	0.6028	0.87
0.02	16.0	912	0.6283	0.85
0.0175	17.0	969	0.6103	0.88
0.0151	18.0	1026	0.6244	0.88
0.014	19.0	1083	0.6293	0.86
0.0181	20.0	1140	0.6333	0.87

Framework versions

• Transformers 4.31.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3