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

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: 0.0297
• Accuracy: 0.9912

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
0.4852	1.0	551	0.4533	0.8042
0.3033	2.0	1102	0.2157	0.9157
0.2339	3.0	1653	0.1212	0.9534
0.1694	4.0	2204	0.1076	0.9603
0.1715	5.0	2755	0.0830	0.9692
0.1339	6.0	3306	0.0674	0.9762
0.1527	7.0	3857	0.0556	0.9791
0.1214	8.0	4408	0.0455	0.9832
0.1062	9.0	4959	0.0466	0.9829
0.0974	10.0	5510	0.0403	0.9849
0.0875	11.0	6061	0.0385	0.9860
0.0992	12.0	6612	0.0376	0.9870
0.065	13.0	7163	0.0392	0.9864
0.0775	14.0	7714	0.0344	0.9890
0.0544	15.0	8265	0.0362	0.9888
0.0584	16.0	8816	0.0422	0.9872
0.0722	17.0	9367	0.0314	0.9900
0.0765	18.0	9918	0.0313	0.9908
0.0696	19.0	10469	0.0297	0.9912
0.0596	20.0	11020	0.0285	0.9910

Framework versions

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