distilhubert-finetuned-gtzan-2

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: 0.9149
• Accuracy: 0.83

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: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0823	1.0	113	2.0903	0.46
1.5111	2.0	226	1.5342	0.6
1.2342	3.0	339	1.1036	0.68
0.8352	4.0	452	0.9137	0.78
0.5727	5.0	565	0.6258	0.81
0.3957	6.0	678	0.5984	0.83
0.1851	7.0	791	0.6269	0.82
0.1607	8.0	904	0.6945	0.79
0.1426	9.0	1017	0.6103	0.86
0.0519	10.0	1130	0.7502	0.81
0.0097	11.0	1243	0.7101	0.85
0.006	12.0	1356	0.8174	0.82
0.0039	13.0	1469	0.8008	0.84
0.0032	14.0	1582	0.8438	0.81
0.0027	15.0	1695	0.8206	0.82
0.0024	16.0	1808	0.8563	0.82
0.002	17.0	1921	0.8884	0.82
0.0018	18.0	2034	0.9148	0.82
0.0018	19.0	2147	0.9017	0.83
0.0016	20.0	2260	0.9178	0.83
0.0015	21.0	2373	0.9070	0.83
0.0014	22.0	2486	0.9033	0.83
0.0014	23.0	2599	0.8975	0.84
0.0013	24.0	2712	0.9160	0.83
0.0013	25.0	2825	0.9149	0.83

Framework versions

• Transformers 4.29.0
• Pytorch 2.0.1
• Datasets 2.12.0
• Tokenizers 0.13.2