This repository is a fork of the original GraphNAS repository. It holds the code for the Evolutionary Algorithm (here) and Random Search (here) strategies, implemented for this paper.
A video presentation can be found here, and the slides can be found here.
Recent versions of PyTorch, numpy, scipy, dgl, and torch_geometric are required.
We have provided a utility script that installs the dependencies, considering the usage of CUDA 10.1. If this is not your CUDA version, follow the instructions on the script.
Example run:
./virtualenv_script.sh /opt/cache # use this parameter if you would like to use a different dir. as pip's cache
After executing this script, you will have an Anaconda powered virtual environment called py37 with the dependencies necessary to run the code in this repository.
We have made available a script for generating the experiment combinations used in the paper. Just run:
./generate_experiment_combinations.sh [ea|rs|rl]
The parameter is the desired optimizer: one of {ea, rl, rs}.
The results are summarized into a jupyter notebook (here). If you would like to re-execute the notebook, please de-compress the results files macro_results.tar.xz and micro_results.tar.xz.
tar -xvf macro_results.tar.xz