convnext-tiny-224_album_vitVMMRdb_make_model_album_pred

This model is a fine-tuned version of facebook/convnext-tiny-224 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7021
• Accuracy: 0.8173
• Precision: 0.8094
• Recall: 0.8173
• F1: 0.8057

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• 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	Precision	Recall	F1
4.6105	1.0	839	4.5248	0.1097	0.0579	0.1097	0.0403
3.4711	2.0	1678	3.3162	0.3000	0.2302	0.3000	0.2097
2.6202	3.0	2517	2.4445	0.4709	0.4120	0.4709	0.3939
2.0614	4.0	3356	1.8839	0.5742	0.5389	0.5742	0.5168
1.7026	5.0	4195	1.5247	0.6436	0.6180	0.6436	0.6013
1.4288	6.0	5034	1.2768	0.6979	0.6810	0.6979	0.6686
1.1953	7.0	5873	1.0960	0.7323	0.7218	0.7323	0.7077
1.058	8.0	6712	0.9828	0.7548	0.7441	0.7548	0.7350
0.9691	9.0	7551	0.9018	0.7718	0.7616	0.7718	0.7536
0.8757	10.0	8390	0.8380	0.7893	0.7806	0.7893	0.7756
0.8446	11.0	9229	0.7905	0.7982	0.7913	0.7982	0.7859
0.7711	12.0	10068	0.7524	0.8069	0.7995	0.8069	0.7950
0.7689	13.0	10907	0.7283	0.8123	0.8043	0.8123	0.8009
0.6919	14.0	11746	0.7133	0.8148	0.8061	0.8148	0.8036
0.694	15.0	12585	0.7064	0.8177	0.8089	0.8177	0.8067

Framework versions

• Transformers 4.24.0
• Pytorch 1.12.1+cu113
• Datasets 2.7.1
• Tokenizers 0.13.2