Skip to content

Latest commit

 

History

History
31 lines (29 loc) · 1.03 KB

README.md

File metadata and controls

31 lines (29 loc) · 1.03 KB
Raw

Uniformly Scaling Flows for Anomaly Detection

This repository contains experiments investigating uniformly scaling normalizing flows for anomaly detection. We use the VeriFlow implementation of uniformly scaling flows.

Installation

  1. Clone the repository
  2. Clone the VeriFlow repository
git clone [email protected]:aai-institute/VeriFlow.git
  1. Create a new virtual environment, e.g. with Conda
conda create -n nf4ad python=3.10
conda activate nf4ad
  1. Install the requirements
pip install path/to/veriflow/project/directory

Run experiments

The experiments for uniformly scaling normalizing flows can be run with the following command from within the nf4ad project directory:

python scripts/run-experiment.py --config experiments/<experiment>/<expconfig>.yaml

where <experiment> is the name of the experiment and <expconfig> is the name of the configuration file for the experiment. The configuration files are located in the experiments directory.