wav2vec2-base-finetuned-gtzan

This model is a fine-tuned version of facebook/wav2vec2-base on the GTZAN dataset. It achieves the following results on the evaluation set:

• Loss: 0.7926
• 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	Validation Loss	Accuracy
2.0468	1.0	113	2.0109	0.41
1.6902	2.0	226	1.6493	0.5
1.0179	3.0	339	1.4098	0.59
1.1239	4.0	452	1.1319	0.67
0.7065	5.0	565	0.9650	0.73
0.546	6.0	678	0.9210	0.75
0.535	7.0	791	0.7329	0.81
0.3793	8.0	904	0.5348	0.86
0.6647	9.0	1017	0.6605	0.84
0.3996	10.0	1130	0.7797	0.83
0.432	11.0	1243	0.7763	0.83
0.0538	12.0	1356	0.7716	0.84
0.0858	13.0	1469	0.7953	0.82
0.3906	14.0	1582	0.7821	0.84
0.2496	15.0	1695	0.9718	0.83
0.13	16.0	1808	0.7773	0.85
0.1103	17.0	1921	0.6670	0.88
0.1443	18.0	2034	0.8843	0.84
0.0083	19.0	2147	0.7977	0.84
0.0086	20.0	2260	0.7926	0.83

Framework versions

• Transformers 4.31.0.dev0
• Pytorch 1.13.0
• Datasets 2.1.0
• Tokenizers 0.13.3