hubert-base-ls960-finetuned-gtzan-efficient

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: 1.0959
• Accuracy: 0.89

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1988	1.0	113	2.1353	0.36
1.6317	2.0	226	1.7387	0.39
1.4411	3.0	339	1.3925	0.46
0.8491	4.0	452	1.0834	0.65
2.1748	5.0	565	1.1530	0.64
1.4915	6.0	678	0.9865	0.69
0.4322	7.0	791	1.3910	0.6
0.6867	8.0	904	1.1252	0.7
0.0758	9.0	1017	0.7395	0.75
1.8782	10.0	1130	0.9792	0.77
1.0492	11.0	1243	0.8810	0.75
0.0376	12.0	1356	0.7031	0.81
0.0648	13.0	1469	0.7527	0.82
1.1951	14.0	1582	0.7731	0.84
0.0071	15.0	1695	0.9237	0.83
0.0095	16.0	1808	0.8471	0.85
0.0014	17.0	1921	1.0585	0.87
0.0007	18.0	2034	1.0959	0.89
0.0003	19.0	2147	1.3957	0.86
3.0069	20.0	2260	1.6382	0.84
0.0	21.0	2373	1.3385	0.88

Framework versions

• Transformers 4.30.2
• Pytorch 2.1.0.dev20230627+cu121
• Datasets 2.13.1
• Tokenizers 0.13.3