hubert-large-ls960-ft-finetuned-gtzan

This model is a fine-tuned version of facebook/hubert-large-ls960-ft on the GTZAN dataset. It achieves the following results on the evaluation set:

• Loss: 1.4835
• Accuracy: 0.76

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
• 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
2.2311	1.0	100	2.2236	0.24
1.9497	2.0	200	1.7922	0.375
1.4897	3.0	300	1.5512	0.4
1.3977	4.0	400	1.5379	0.455
1.0858	5.0	500	1.4778	0.535
1.3193	6.0	600	1.1541	0.59
0.9246	7.0	700	1.3068	0.595
0.8115	8.0	800	1.0093	0.67
0.7293	9.0	900	1.1365	0.67
0.7645	10.0	1000	1.0879	0.69
0.6447	11.0	1100	1.1747	0.69
0.2322	12.0	1200	1.0627	0.73
0.2428	13.0	1300	0.9681	0.765
0.2777	14.0	1400	1.3665	0.72
0.2792	15.0	1500	1.3216	0.73
0.2509	16.0	1600	1.2809	0.755
0.7852	17.0	1700	1.3793	0.77
0.3948	18.0	1800	1.4736	0.765
0.3591	19.0	1900	1.5412	0.76
0.0059	20.0	2000	1.4835	0.76

Framework versions

• Transformers 4.50.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.3.1
• Tokenizers 0.21.0