wav2vec2-base-music_genre_classifier-g4-firstseconds

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

• Loss: 1.4970
• Accuracy: 0.8304
• F1: 0.8244
• Recall: 0.8262
• Precision: 0.8260

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: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• 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	F1	Recall	Precision
1.1447	1.0	2393	1.2684	0.6524	0.6382	0.6533	0.6527
0.7302	2.0	4786	0.8886	0.7458	0.7421	0.7479	0.7637
0.521	3.0	7179	0.8755	0.7701	0.7686	0.7715	0.7934
0.3648	4.0	9572	1.0389	0.7731	0.7723	0.7673	0.7928
0.6132	5.0	11965	1.0694	0.7997	0.7955	0.7943	0.8170
0.6512	6.0	14358	1.2190	0.7886	0.7864	0.7864	0.7984
0.0851	7.0	16751	1.2496	0.8022	0.7959	0.7973	0.8082
0.0881	8.0	19144	1.2582	0.8127	0.8088	0.8105	0.8098
0.1063	9.0	21537	1.4087	0.8148	0.8119	0.8121	0.8176
0.4205	10.0	23930	1.4825	0.8055	0.8001	0.8019	0.8158
0.0478	11.0	26323	1.4240	0.8109	0.8023	0.8031	0.8082
0.0037	12.0	28716	1.3865	0.8248	0.8182	0.8199	0.8202
0.0236	13.0	31109	1.4570	0.8279	0.8230	0.8232	0.8250
0.0094	14.0	33502	1.4892	0.8289	0.8227	0.8248	0.8249
0.0002	15.0	35895	1.4970	0.8304	0.8244	0.8262	0.8260

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.0+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3