distil-ast-audioset-finetuned-gtzan

This model is a fine-tuned version of bookbot/distil-ast-audioset on the GTZAN dataset. It achieves the following results on the evaluation set:

• Loss: 0.3866
• Accuracy: 0.9

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: 1e-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
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
0.7945	1.0	113	0.76	0.8481
0.4836	2.0	226	0.79	0.5647
0.2434	3.0	339	0.8	0.6345
0.2721	4.0	452	0.91	0.3684
0.0459	5.0	565	0.9	0.3387
0.0485	6.0	678	0.87	0.3720
0.0337	7.0	791	0.9	0.3439
0.0206	8.0	904	0.89	0.3630
0.1043	9.0	1017	0.89	0.3682
0.0146	10.0	1130	0.89	0.3552
0.0109	11.0	1243	0.4141	0.87
0.001	12.0	1356	0.4266	0.88
0.0006	13.0	1469	0.4001	0.91
0.0006	14.0	1582	0.3884	0.9
0.0008	15.0	1695	0.3881	0.91
0.0005	16.0	1808	0.3796	0.9
0.0005	17.0	1921	0.3865	0.9
0.0005	18.0	2034	0.3861	0.9
0.0003	19.0	2147	0.3879	0.9
0.0005	20.0	2260	0.3866	0.9

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3