senthilsdglakhsg/distilhubert-gtzan-15sec

distilhubert-finetuned-gtzan-15sec

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: 0.8519
• Accuracy: 0.8492

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: 0.0001
• 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_steps: 100
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
3.1339	1.0	112	1.3512	0.6583
1.6185	2.0	224	0.8620	0.7085
1.0020	3.0	336	0.5799	0.8090
0.7062	4.0	448	0.6624	0.8040
0.5191	5.0	560	0.6755	0.8191
0.2967	6.0	672	0.7785	0.7990
0.0821	7.0	784	0.8235	0.8342
0.0686	8.0	896	0.8519	0.8492
0.0156	9.0	1008	0.9783	0.8191
0.0019	10.0	1120	1.0472	0.8191
0.0015	11.0	1232	1.1903	0.8090
0.0862	12.0	1344	1.1610	0.8191
0.0011	13.0	1456	1.1986	0.8191
0.0008	14.0	1568	1.2203	0.8141
0.0010	15.0	1680	1.4290	0.7889
0.0007	16.0	1792	1.3051	0.8090
0.0007	17.0	1904	1.1839	0.8291
0.0077	18.0	2016	1.1803	0.8342
0.0007	19.0	2128	1.1789	0.8291
0.0006	20.0	2240	1.1757	0.8342

Framework versions

• Transformers 5.16.1
• Pytorch 2.11.0+cu128
• Datasets 5.0.1
• Tokenizers 0.23.1