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: 0.7650
• Accuracy: 0.88

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: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2258	1.0	225	1.9240	0.28
1.6083	2.0	450	1.4887	0.39
1.3983	3.0	675	1.3524	0.56
0.7368	4.0	900	1.3110	0.56
0.6121	5.0	1125	0.9572	0.72
0.1772	6.0	1350	0.8775	0.73
1.8666	7.0	1575	0.6078	0.82
0.091	8.0	1800	0.9999	0.76
0.0458	9.0	2025	0.7169	0.83
0.6817	10.0	2250	0.7614	0.86
0.7023	11.0	2475	0.9348	0.84
0.0047	12.0	2700	0.7222	0.88
0.0363	13.0	2925	0.7027	0.89
0.0073	14.0	3150	0.7440	0.88
0.0055	15.0	3375	0.7650	0.88

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0