ExaRL on Darwin (at LANL)

Get source code

git clone --recursive https://github.com/exalearn/EXARL.git
cd EXARL
git checkout develop # assuming you want the latest development version

Set up environment

module load miniconda3
conda create --name exarl
source activate exarl
conda install -c anaconda tensorflow-gpu
conda install -c conda-forge mpi4py gym ase lmfit scikit-learn pandas numba pybind11 pytest
pip install plotille
pip install bsuite
export PYTHONPATH=<your path to EXARL>:$PYTHONPATH

Thereafter, once you get a compute node do this setup for your environment

module load miniconda3 openmpi
source activate exarl
export PYTHONPATH=<your path to EXARL>:$PYTHONPATH

This runs a bsuite test seed

python ./exarl/driver --env Bsuite-v0 --bsuite_id <name of bsuite test> --seed_number 0 --n_episodes 1000 --n_steps 1000 --output_dir <your test output dir>