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: 2.2594
• Accuracy: 0.82

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: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2669	1.0	57	2.2222	0.29
1.9365	2.0	114	1.8485	0.53
1.5115	3.0	171	1.4544	0.64
1.1314	4.0	228	1.1404	0.7
0.9473	5.0	285	0.9750	0.7
0.8026	6.0	342	0.8381	0.76
0.669	7.0	399	0.7231	0.81
0.5026	8.0	456	0.7019	0.8
0.3179	9.0	513	0.6318	0.81
0.2934	10.0	570	0.6551	0.81
0.1709	11.0	627	0.6041	0.81
0.1502	12.0	684	0.7066	0.84
0.0626	13.0	741	0.6859	0.84
0.0184	14.0	798	0.7444	0.8
0.0345	15.0	855	0.9701	0.8
0.0034	16.0	912	1.0236	0.83
0.0014	17.0	969	1.1226	0.81
0.0811	18.0	1026	1.2570	0.81
0.0002	19.0	1083	1.3850	0.81
0.0	20.0	1140	1.6715	0.82
0.0	21.0	1197	1.8665	0.8
0.1033	22.0	1254	1.8919	0.79
0.047	23.0	1311	1.9730	0.82
0.0	24.0	1368	2.1126	0.81
0.0	25.0	1425	2.1545	0.79
0.0	26.0	1482	2.2609	0.79
0.0	27.0	1539	2.2284	0.81
0.0	28.0	1596	2.2374	0.81
0.0	29.0	1653	2.2590	0.82
0.0	30.0	1710	2.2594	0.82

Framework versions

• Transformers 4.43.0.dev0
• Pytorch 2.3.1+cu118
• Datasets 2.20.0
• Tokenizers 0.19.1