This project is an unofficial implementation of "EfficientAD: Accurate Visual Anomaly Detection at Millisecond-Level Latencies", which is implemented step-by-step according to the pseudocode in the appendix
./data
-
ImageNet
- n01440764
- n01443537 ...
-
MVTec_AD
- bottle
- ground_truth
- test
- train
- cable
- ground_truth
- test
- train ...
- bottle
-
result
conda activate <your_env>
pip install -r requirements.txt
python distillaion_training.py
python train_reduced_student.py
This repository is implemented according to the pseudocode in the appendix, which may differ from the official version and has not been tested for performance. It does not represent the true effectiveness of EfficientAD. If you have any suggestions, please submit an issue or PR.