hubert-base-ls960-finetuned-gtzan

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

• Loss: 1.0283
• Accuracy: 0.83

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
2.2494	1.0	113	0.36	2.1568
1.7795	2.0	226	0.38	1.7904
1.5798	3.0	339	0.5	1.6144
1.6354	4.0	452	0.66	1.2584
0.9675	5.0	565	0.64	1.1453
0.995	6.0	678	0.67	0.9740
1.2052	7.0	791	0.68	1.0552
0.7028	8.0	904	0.74	0.8980
0.7472	9.0	1017	0.72	0.9431
0.3181	10.0	1130	0.75	0.8750
0.3948	11.0	1243	0.73	1.0047
0.3507	12.0	1356	0.81	0.8054
0.1785	13.0	1469	0.84	0.7866
0.2453	14.0	1582	0.82	0.8960
0.2832	15.0	1695	0.81	1.0770
0.2132	16.0	1808	0.82	0.9359
0.1398	17.0	1921	0.81	1.0800
0.292	18.0	2034	0.84	0.9867
0.0181	19.0	2147	0.82	1.0585
0.0399	20.0	2260	1.0283	0.83

Framework versions

• Transformers 4.34.1
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1