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: 0.7472
• Accuracy: 0.81

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: 3e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• 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	Accuracy	Validation Loss
2.2042	1.0	112	0.27	2.1274
1.7875	2.0	225	0.51	1.6840
1.4927	3.0	337	0.57	1.3809
1.2344	4.0	450	0.64	1.2021
1.2579	5.0	562	0.62	1.1646
0.9661	6.0	675	0.65	1.0412
1.0119	7.0	787	0.74	0.8671
0.8629	8.0	900	0.66	0.9364
0.607	9.0	1012	0.75	0.8867
0.5699	10.0	1125	0.78	0.7432
0.5128	11.0	1237	0.76	0.8212
0.4203	12.0	1350	0.77	0.8128
0.348	13.0	1462	0.81	0.7472
0.3869	14.0	1575	0.8	0.7456
0.2129	14.93	1680	0.79	0.7243

Framework versions

• Transformers 4.38.2
• Pytorch 2.2.0
• Datasets 2.18.0
• Tokenizers 0.15.2