Installation Instructions

In a conda environment or virtual environment with Python 3.8, change directories into /path/to/satnet_repo and run:

pip install -e .

To run training:

Activate your conda or virtual environment, change directories to /path/to/satnet_repo/satnet/scripts, and run:

python train.py

During training, you can view the agent's learning progress in tensorboard by running (in another terminal):

tensorboard --logdir=/path/to/ray_results

where you can specify this path in train.py under the results_dir variable.

To run inference/rollout:

Coming soon!

References:

1. IEEE Aerospace Conference paper: https://ieeexplore.ieee.org/abstract/document/9438519/
2. AAAI ML4OR Workshop paper: https://openreview.net/forum?id=buIUxK7F-Bx