Skip to content

taikiinoue45/RIAD

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

30 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

black blackdoc flake8 isort mypy

PyTorch re-implementation of Reconstruction by Inpainting for Visual Anomaly Detection


1. AUROC Scores

category Paper My Implementation
zipper 0.981 0.975
wood 0.930 0.965
transistor 0.909 0.918
toothbrush 1.000 0.972
tile 0.987 0.997
screw 0.845 0.799
pill 0.838 0.786
metal_nut 0.885 0.920
leather 1.000 1.000
hazelnut 0.833 0.890
grid 0.996 0.983
carpet 0.842 0.781
capsule 0.884 0.731
cable 0.819 0.655
bottle 0.999 0.971

2. Graphical Results

zipper

wood

transistor

toothbrush

tile

screw

pill

metal_nut

leather

hazelnut

grid

carpet

capsule

cable

bottle


3. Requirements

  • CUDA 10.2
  • nvidia-docker2

4. Usage

a) Download docker image and run docker container

docker pull taikiinoue45/mvtec:riad
docker run --runtime nvidia -it --workdir /app --network host taikiinoue45/mvtec:riad /usr/bin/zsh

b) Download this repository

git clone https://github.com/taikiinoue45/RIAD.git
cd /app/RIAD/riad

c) Run experiments

sh run.sh

d) Visualize experiments

mlflow ui

5. Contacts