vit-base-patch16-224-finetuned-main-gpu-20e-final-1

This model is a fine-tuned version of google/vit-base-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.0272
• Accuracy: 0.9918

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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
0.4776	1.0	551	0.4399	0.8125
0.3207	2.0	1102	0.2645	0.8978
0.2292	3.0	1653	0.1388	0.9468
0.1811	4.0	2204	0.0943	0.9662
0.1633	5.0	2755	0.0740	0.9723
0.1355	6.0	3306	0.0744	0.9727
0.1413	7.0	3857	0.0548	0.9813
0.1257	8.0	4408	0.0442	0.9844
0.1057	9.0	4959	0.0517	0.9821
0.1	10.0	5510	0.0376	0.9868
0.0873	11.0	6061	0.0410	0.9866
0.0974	12.0	6612	0.0430	0.9861
0.0673	13.0	7163	0.0421	0.9852
0.0913	14.0	7714	0.0339	0.9882
0.0594	15.0	8265	0.0327	0.9896
0.0608	16.0	8816	0.0379	0.9885
0.0725	17.0	9367	0.0288	0.9904
0.0742	18.0	9918	0.0284	0.9906
0.0708	19.0	10469	0.0273	0.9916
0.0648	20.0	11020	0.0272	0.9918

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2