atharva98/distilhubert-finetuned-gtzan-increased-accuracy

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: 0.6492
• 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: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use adamw_torch 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
2.2305	1.0	200	2.1948	0.37
1.8318	2.0	400	1.7088	0.555
1.44	3.0	600	1.3910	0.565
1.0672	4.0	800	1.0886	0.725
0.6987	5.0	1000	1.0055	0.72
0.6847	6.0	1200	0.8528	0.755
0.8731	7.0	1400	0.7888	0.78
0.7602	8.0	1600	0.6399	0.81
0.4625	9.0	1800	0.6973	0.795
0.1884	10.0	2000	0.6347	0.815
0.0945	11.0	2200	0.6723	0.805
0.0334	12.0	2400	0.7025	0.815
0.2925	13.0	2600	0.5688	0.835
0.0322	14.0	2800	0.5742	0.845
0.0166	15.0	3000	0.5655	0.86
0.1173	16.0	3200	0.5723	0.855
0.0223	17.0	3400	0.6601	0.83
0.0103	18.0	3600	0.6294	0.835
0.0096	19.0	3800	0.6413	0.835
0.0093	20.0	4000	0.6492	0.84

Framework versions

• Transformers 4.52.0.dev0
• Pytorch 2.8.0.dev20250404+cu128
• Datasets 3.5.0
• Tokenizers 0.21.1