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: 1.1749
• Accuracy: 0.86

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: 4
• eval_batch_size: 4
• 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
2.0517	1.0	225	2.0004	0.47
1.3283	2.0	450	1.3458	0.57
0.729	3.0	675	0.8563	0.76
0.4007	4.0	900	0.6748	0.8
0.3923	5.0	1125	0.7340	0.78
0.2193	6.0	1350	0.8712	0.76
0.2383	7.0	1575	0.7414	0.79
0.3	8.0	1800	0.7387	0.86
0.006	9.0	2025	0.9203	0.85
0.002	10.0	2250	0.8956	0.85
0.0014	11.0	2475	0.9831	0.86
0.001	12.0	2700	0.9406	0.86
0.0009	13.0	2925	1.0288	0.86
0.0007	14.0	3150	1.0172	0.86
0.0007	15.0	3375	0.9912	0.89
0.0005	16.0	3600	1.0282	0.86
0.0006	17.0	3825	1.3495	0.83
0.2453	18.0	4050	1.0340	0.87
0.0004	19.0	4275	1.1048	0.86
0.0004	20.0	4500	1.3051	0.85
0.0003	21.0	4725	1.2280	0.85
0.0003	22.0	4950	1.2530	0.85
0.0003	23.0	5175	1.1992	0.85
0.0003	24.0	5400	1.1881	0.85
0.0003	25.0	5625	1.1749	0.86

Framework versions

• Transformers 4.44.2
• Pytorch 2.3.1
• Datasets 2.21.0
• Tokenizers 0.19.1