vit-base-patch16-224-in21k-bridgedefectVIT15

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2402
• Accuracy: {'accuracy': 0.9573153608536927}
• F1: {'f1': 0.9566147291413047}
• Precision: {'precision': 0.9591127716274309}
• Recall: {'recall': 0.9565472623176632}

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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	F1	Precision	Recall
0.3548	1.0	1780	0.2848	{'accuracy': 0.9118225217635496}	{'f1': 0.912598515170384}	{'precision': 0.913326374297146}	{'recall': 0.9157022464716918}
0.1718	2.0	3560	0.3435	{'accuracy': 0.9005897219882055}	{'f1': 0.9021520907258462}	{'precision': 0.9071588887385811}	{'recall': 0.9088734326741875}
0.1956	3.0	5340	0.2290	{'accuracy': 0.9337264813254704}	{'f1': 0.9345043308561282}	{'precision': 0.9371641968965463}	{'recall': 0.9353444695340449}
0.1589	4.0	7120	0.3518	{'accuracy': 0.925582701488346}	{'f1': 0.9240312800580016}	{'precision': 0.9310407182465765}	{'recall': 0.9241275251443595}
0.1076	5.0	8900	0.4017	{'accuracy': 0.9188430216231396}	{'f1': 0.9170326424426785}	{'precision': 0.923800610078333}	{'recall': 0.9181896594596475}
0.0895	6.0	10680	0.2950	{'accuracy': 0.938219601235608}	{'f1': 0.9380460882172743}	{'precision': 0.9406510771971466}	{'recall': 0.9398150744796098}
0.0833	7.0	12460	0.1882	{'accuracy': 0.9559112608817748}	{'f1': 0.9553785330080078}	{'precision': 0.957564211420095}	{'recall': 0.9550045684543612}
0.034	8.0	14240	0.3222	{'accuracy': 0.9401853411962932}	{'f1': 0.9401162584753809}	{'precision': 0.944463542451817}	{'recall': 0.9410746120960137}
0.1117	9.0	16020	0.3084	{'accuracy': 0.9401853411962932}	{'f1': 0.9389336455514373}	{'precision': 0.945493350000876}	{'recall': 0.9374486305327216}
0.2057	10.0	17800	0.3612	{'accuracy': 0.9348497613030048}	{'f1': 0.9343390020827073}	{'precision': 0.939876035403298}	{'recall': 0.9348316142752356}
0.1	11.0	19580	0.2284	{'accuracy': 0.9553496208930076}	{'f1': 0.9540937018628736}	{'precision': 0.9563364479044711}	{'recall': 0.9537814730817218}
0.0531	12.0	21360	0.2393	{'accuracy': 0.9528222409435552}	{'f1': 0.9517895350619009}	{'precision': 0.955245168398952}	{'recall': 0.9514588091149371}
0.0597	13.0	23140	0.2695	{'accuracy': 0.9519797809604044}	{'f1': 0.9513321647748849}	{'precision': 0.9541412213348108}	{'recall': 0.9515688542696423}
0.0482	14.0	24920	0.2403	{'accuracy': 0.9567537208649256}	{'f1': 0.9560207781245073}	{'precision': 0.9590114685856663}	{'recall': 0.9557731012948057}
0.0019	15.0	26700	0.2402	{'accuracy': 0.9573153608536927}	{'f1': 0.9566147291413047}	{'precision': 0.9591127716274309}	{'recall': 0.9565472623176632}

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0
• Datasets 2.17.1
• Tokenizers 0.15.2