Anomaly Detection through Conditional Diffusion Probability Modeling on Graphs

This repository is the official implementation of CGADM.

Requirements

To install requirements:

conda env create -f gad.yaml

Data Download

Download all the datasets from the following links and put them in the "dataset" folder: Download datasets

Training and Evaluation

To train the model(s) in the paper:

python main.py

Output: the file "best_XXXXXX.tar" in the "data" folder