hubert-base-ls960-finetuned-gtzan

This model is a fine-tuned version of facebook/hubert-base-ls960 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1107
• Accuracy: 0.85

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
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
1.9975	1.0	225	0.47	1.8130
1.2415	2.0	450	0.57	1.3022
1.0225	3.0	675	0.645	1.1478
1.1012	4.0	900	0.755	0.8725
1.0753	5.0	1125	0.67	1.1645
0.5354	6.0	1350	0.66	1.3094
0.7805	7.0	1575	0.795	0.8406
0.3307	8.0	1800	0.795	0.9782
0.1861	9.0	2025	0.79	0.9140
0.2776	10.0	2250	0.795	1.1711
0.314	11.0	2475	0.825	0.9193
0.1785	12.0	2700	0.82	1.0272
0.1444	13.0	2925	0.845	0.9903
0.0122	14.0	3150	0.835	0.9974
0.0116	15.0	3375	0.85	0.9670
0.3403	31.0	3472	1.0085	0.85
0.3596	32.0	3585	1.3101	0.81
0.0242	33.0	3697	0.9612	0.86
0.1006	34.0	3810	1.1904	0.82
0.1034	35.0	3922	0.9582	0.86
0.195	36.0	4035	1.0223	0.84
0.0081	37.0	4147	1.2461	0.8
0.006	38.0	4260	0.9541	0.87
0.281	39.0	4372	0.9340	0.87
0.0491	40.0	4485	1.0942	0.85
0.0537	41.0	4597	1.1521	0.85
0.0017	42.0	4710	1.1738	0.85
0.0031	43.0	4822	1.1584	0.85
0.1107	44.0	4935	1.1503	0.86
0.0032	45.0	5047	1.0710	0.87
0.0027	46.0	5160	1.1310	0.86
0.0013	47.0	5272	1.1194	0.86
0.0023	48.0	5385	1.1173	0.85
0.0286	49.0	5497	1.1087	0.85
0.0133	49.91	5600	1.1107	0.85

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1
• Datasets 2.15.0
• Tokenizers 0.13.2