vit-base-patch16-224_album_vitVMMRdb_make_model_album_pred

This model is a fine-tuned version of google/vit-base-patch16-224 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4670
• Accuracy: 0.8781
• Precision: 0.8768
• Recall: 0.8781
• F1: 0.8758

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
3.5529	1.0	839	3.3687	0.3096	0.2809	0.3096	0.2246
1.7855	2.0	1678	1.6042	0.6378	0.6187	0.6378	0.5996
1.1054	3.0	2517	1.0105	0.7556	0.7512	0.7556	0.7385
0.8179	4.0	3356	0.7794	0.8033	0.8020	0.8033	0.7934
0.6057	5.0	4195	0.6479	0.8294	0.8274	0.8294	0.8212
0.4709	6.0	5034	0.5817	0.8478	0.8477	0.8478	0.8428
0.3962	7.0	5873	0.5333	0.8571	0.8570	0.8571	0.8527
0.346	8.0	6712	0.5073	0.8638	0.8647	0.8638	0.8615
0.2772	9.0	7551	0.4881	0.8681	0.8679	0.8681	0.8656
0.2136	10.0	8390	0.4777	0.8719	0.8718	0.8719	0.8689
0.1937	11.0	9229	0.4737	0.8734	0.8731	0.8734	0.8703
0.1754	12.0	10068	0.4604	0.8758	0.8750	0.8758	0.8733
0.1111	13.0	10907	0.4561	0.8790	0.8782	0.8790	0.8768
0.1128	14.0	11746	0.4519	0.8808	0.8799	0.8808	0.8787
0.1018	15.0	12585	0.4497	0.8813	0.8805	0.8813	0.8794

Framework versions

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