levit-128-finetuned-flower

This model is a fine-tuned version of facebook/levit-128 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.1807
• Accuracy: 0.9506
• Precision: 0.9510
• Recall: 0.9506
• F1: 0.9506

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: 0.005
• 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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.6679	1.0	40	0.6957	0.8076	0.8492	0.8076	0.8060
0.7188	2.0	80	0.7094	0.7822	0.7997	0.7822	0.7789
0.7277	3.0	120	0.7803	0.7477	0.7912	0.7477	0.7480
0.561	4.0	160	0.5489	0.8352	0.8462	0.8352	0.8292
0.4958	5.0	200	0.4067	0.8770	0.8852	0.8770	0.8766
0.4681	6.0	240	0.4801	0.8457	0.8570	0.8457	0.8423
0.368	7.0	280	0.4348	0.8617	0.8697	0.8617	0.8618
0.355	8.0	320	0.3401	0.8926	0.8971	0.8926	0.8924
0.3164	9.0	360	0.3510	0.8871	0.8935	0.8871	0.8871
0.2972	10.0	400	0.2877	0.9140	0.9159	0.9140	0.9133
0.2639	11.0	440	0.2588	0.9245	0.9246	0.9245	0.9233
0.264	12.0	480	0.2811	0.9096	0.9155	0.9096	0.9097
0.2082	13.0	520	0.2368	0.9238	0.9244	0.9238	0.9225
0.1506	14.0	560	0.2552	0.9205	0.9244	0.9205	0.9200
0.179	15.0	600	0.2133	0.9401	0.9421	0.9401	0.9399
0.1388	16.0	640	0.2170	0.9376	0.9388	0.9376	0.9377
0.116	17.0	680	0.1817	0.9466	0.9468	0.9466	0.9464
0.0976	18.0	720	0.1915	0.9470	0.9477	0.9470	0.9473
0.0806	19.0	760	0.1876	0.9492	0.9501	0.9492	0.9493
0.0911	20.0	800	0.1807	0.9506	0.9510	0.9506	0.9506

Framework versions

• Transformers 4.39.3
• Pytorch 2.0.1
• Datasets 2.18.0
• Tokenizers 0.15.2