wav2vec2-base-finetuned-gtzan-bs-16

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: 0.5497
• 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: 16
• eval_batch_size: 16
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0557	1.0	57	1.9783	0.34
1.6173	2.0	114	1.6407	0.55
1.3884	3.0	171	1.2228	0.65
1.1082	4.0	228	1.0989	0.7
0.9112	5.0	285	0.8724	0.8
0.7985	6.0	342	0.8715	0.76
0.5456	7.0	399	0.6832	0.82
0.4842	8.0	456	0.6566	0.85
0.3419	9.0	513	0.6485	0.84
0.5821	10.0	570	0.5636	0.85
0.2112	11.0	627	0.4572	0.89
0.2005	12.0	684	0.5405	0.87
0.1314	13.0	741	0.4695	0.9
0.0866	14.0	798	0.5545	0.88
0.0594	15.0	855	0.5497	0.88

Framework versions

• Transformers 4.32.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.3
• Tokenizers 0.13.3