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: 0.7299
• 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
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• 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
2.251	1.0	57	2.1816	0.45
1.7315	2.0	114	1.6342	0.53
1.3194	3.0	171	1.2904	0.65
0.9783	4.0	228	1.0165	0.72
0.8122	5.0	285	0.8711	0.8
0.669	6.0	342	0.7628	0.74
0.5481	7.0	399	0.6805	0.81
0.3229	8.0	456	0.7178	0.78
0.2907	9.0	513	0.6567	0.81
0.2137	10.0	570	0.6404	0.81
0.132	11.0	627	0.6389	0.79
0.0763	12.0	684	0.6886	0.81
0.0483	13.0	741	0.6255	0.84
0.0363	14.0	798	0.6986	0.82
0.0253	15.0	855	0.6512	0.83
0.0203	16.0	912	0.6776	0.83
0.0177	17.0	969	0.7469	0.83
0.0483	18.0	1026	0.7146	0.82
0.0151	19.0	1083	0.7323	0.83
0.0148	20.0	1140	0.7299	0.83

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 3.6.0
• Tokenizers 0.22.0