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: 1.7615
• Accuracy: 0.85

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: 16
• 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: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2914	1.0	57	2.2595	0.25
2.1225	2.0	114	2.0265	0.57
1.7631	3.0	171	1.6482	0.59
1.3445	4.0	228	1.3380	0.62
1.1548	5.0	285	1.0589	0.72
0.9289	6.0	342	0.8541	0.76
0.708	7.0	399	0.7628	0.79
0.4497	8.0	456	0.7088	0.82
0.4061	9.0	513	0.6118	0.85
0.286	10.0	570	0.6684	0.79
0.1739	11.0	627	0.5965	0.83
0.1103	12.0	684	0.8414	0.81
0.0922	13.0	741	0.5937	0.87
0.0166	14.0	798	0.5786	0.86
0.0075	15.0	855	0.7950	0.84
0.0014	16.0	912	0.8492	0.87
0.0006	17.0	969	1.2642	0.82
0.0815	18.0	1026	1.1173	0.87
0.0	19.0	1083	1.2181	0.86
0.0	20.0	1140	1.6673	0.85
0.0	21.0	1197	1.4749	0.86
0.0611	22.0	1254	2.2533	0.82
0.0978	23.0	1311	2.0092	0.86
0.0	24.0	1368	2.3586	0.83
0.0	25.0	1425	1.7617	0.86
0.0	26.0	1482	1.7425	0.86
0.0	27.0	1539	1.8418	0.85
0.0	28.0	1596	1.6987	0.87
0.0	29.0	1653	1.9399	0.85
0.0	30.0	1710	2.4230	0.81
0.0	31.0	1767	1.4312	0.88
0.1807	32.0	1824	1.5278	0.87
0.0	33.0	1881	1.3795	0.88
0.0	34.0	1938	1.5051	0.88
0.0	35.0	1995	1.6587	0.85
0.0	36.0	2052	1.6256	0.86
0.0	37.0	2109	1.7290	0.85
0.0	38.0	2166	1.8676	0.87
0.0	39.0	2223	1.8963	0.86
0.166	40.0	2280	1.7057	0.85
0.1293	41.0	2337	1.4235	0.87
0.1491	42.0	2394	1.7916	0.85
0.1416	43.0	2451	1.8634	0.85
0.0	44.0	2508	1.6286	0.86
0.0526	45.0	2565	1.6242	0.86
0.0	46.0	2622	1.7576	0.85
0.0	47.0	2679	1.7897	0.85
0.0	48.0	2736	1.7571	0.85
0.0018	49.0	2793	1.6993	0.85
0.0	50.0	2850	1.7615	0.85

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1