This is the main repository providing the first non-cohesive framework for General Explainable and Verifiable Artificial Intelligence. Please look at this wiki for more information on this project.
Ensure you have /python/src as your root directory:
export PYTHONPATH="${PYTHONPATH}:/python/src"Install Poetry, if you haven't already:
curl -sSL https://install.python-poetry.org | python3 -Then:
pip install .examples/RunExperiment.py contains example pipeline.
Once results are collected, specify folder with all desired outputs (/results) within: python/src/GEVAI/benchmarking/aggregate_metrics.py, and run the script. A file (output.csv) will be presented comparing: accuracy, F1 score, precision, recall.
Use python/src/GEVAI/benchmarking/plot_results.py to plot how the different results compare in terms of running times, this file should be generated within results/ as benchmark_(x)ex_post_explainers.csv.
Must have CUDA version >= 12.3 installed