The idea of ArgusEyes is to use mlinspect to extract the intermediate results and their lineage from native ML pipelines, and enable a set of automated issue detection and data refinement techniques.
We provide a couple of example pipelines and configurations to showcase the usage of ArgusEyes (note that you can run mlflow ui --backend-store-uri ./mlruns from the project root to view the captured artifacts and issue detection results).
Furthermore, we setup an example project which demonstrates how to integrate ArgusEyes into a github CI workflow.
Prerequisite: Python 3.9
-
Clone this repository
-
Set up the environment
cd arguseyes
python -m venv venv
source venv/bin/activate -
Install graphviz
Linux:apt-get install graphviz
MAC OS:brew install graphviz -
Install pip dependencies
pip install -r requirements.txt