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

Loss: 0.3193
Accuracy: 0.9286
Defect Precision: 0.9216
Defect Recall: 1.0
Defect F1: 0.9592

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: 16
eval_batch_size: 16
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Defect Precision	Defect Recall	Defect F1
0.4932	1.0	17	0.4328	0.8393	0.8393	1.0	0.9126
0.4970	2.0	34	0.4307	0.8393	0.8393	1.0	0.9126
0.3777	3.0	51	0.3795	0.8393	0.8393	1.0	0.9126
0.3863	4.0	68	0.3475	0.8393	0.8393	1.0	0.9126
0.3000	5.0	85	0.3285	0.8393	0.9318	0.8723	0.9011
0.3114	6.0	102	0.4423	0.8393	0.8393	1.0	0.9126
0.2635	7.0	119	0.3369	0.8571	0.9333	0.8936	0.9130
0.2046	8.0	136	0.4730	0.7679	0.925	0.7872	0.8506
0.2625	9.0	153	0.3185	0.8929	0.9362	0.9362	0.9362
0.2306	10.0	170	0.3193	0.9286	0.9216	1.0	0.9592

Framework versions

Transformers 5.12.1
Pytorch 2.12.1+cu130
Datasets 5.0.0
Tokenizers 0.22.2

Dongjin1203/defect-classifier-vit-base

defect-classifier-vit-base

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Dongjin1203/
defect-classifier-vit-base