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: 1.2162
• Accuracy: 0.84

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: 6e-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: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2216	1.0	57	2.1520	0.45
1.704	2.0	114	1.6254	0.59
1.1885	3.0	171	1.1741	0.7
0.8575	4.0	228	0.9653	0.7
0.6822	5.0	285	0.8219	0.81
0.5103	6.0	342	0.7254	0.8
0.4386	7.0	399	0.6772	0.84
0.2862	8.0	456	0.7047	0.8
0.1639	9.0	513	0.7126	0.8
0.0998	10.0	570	0.8339	0.77
0.0585	11.0	627	0.7380	0.8
0.0256	12.0	684	0.7606	0.84
0.0201	13.0	741	0.8292	0.83
0.0058	14.0	798	0.9495	0.83
0.0029	15.0	855	1.1009	0.82
0.0016	16.0	912	1.1451	0.82
0.001	17.0	969	1.1886	0.84
0.0071	18.0	1026	1.1731	0.84
0.0004	19.0	1083	1.2255	0.84
0.0003	20.0	1140	1.2162	0.84

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.19.1