distilhubert-finetuned-gtzan-30-epochs

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: 1.1939
• Accuracy: 0.81

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1804	1.0	113	2.1756	0.46
1.7271	2.0	226	1.6973	0.53
1.2703	3.0	339	1.2950	0.51
0.9446	4.0	452	0.9433	0.68
0.6192	5.0	565	0.7885	0.73
0.3628	6.0	678	0.8338	0.76
0.2871	7.0	791	0.8125	0.74
0.0587	8.0	904	0.7500	0.8
0.1316	9.0	1017	0.8711	0.79
0.0175	10.0	1130	0.7429	0.82
0.0818	11.0	1243	0.9848	0.81
0.0049	12.0	1356	1.0498	0.76
0.0034	13.0	1469	1.0422	0.84
0.0028	14.0	1582	1.0919	0.83
0.0023	15.0	1695	1.0565	0.82
0.0019	16.0	1808	1.0797	0.84
0.0769	17.0	1921	1.1430	0.82
0.104	18.0	2034	1.1482	0.8
0.0014	19.0	2147	1.0972	0.83
0.0012	20.0	2260	1.1867	0.82
0.0012	21.0	2373	1.1914	0.82
0.0012	22.0	2486	1.1461	0.84
0.0009	23.0	2599	1.1401	0.82
0.0009	24.0	2712	1.1686	0.84
0.0009	25.0	2825	1.1824	0.85
0.0009	26.0	2938	1.1815	0.81
0.0008	27.0	3051	1.1808	0.82
0.0008	28.0	3164	1.1904	0.81
0.0008	29.0	3277	1.1990	0.82
0.0008	30.0	3390	1.1939	0.81

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3