glmcdona/LuxPythonEnvGym: Matching python environment code for Lux AI 2021 Kaggle competition, and a gym interface for RL models.

Matching python environment code for Lux AI 2021 Kaggle competition, and a gym interface for RL models.

This is still in-progress and not ready for use yet. I’ll post official docs and usage guide here once a simple example converges, and can be transformed to a kaggle-submission format.

Installation

See above, this project isn’t complete yet. If you want to try your own agents anyways, here are instructions. This should work cross-platform, but I’ve only tested Windows 10 and Ubuntu.

Install luxai2021 environment package by running the installer:

python setup.py install

Now edit and run your ML agent training code:

python ./examples/train.py

You can then run tensorboard to monitor the training:

tensorboard --logdir lux_tensorboard

Example kaggle notebook

www.kaggle.com/glmcdona/lux-ai-deep-reinforcement-learning-ppo-example

Preparing a kaggle submission

You have trained a model, and now you’d like to submit it as a kaggle submission. Here are the steps to prepare your submission.

Either view the above kaggle example or prepare a submission yourself:

1. Place your trained model file as model.zip and your agent file agent_policy.py in the ./kaggle_submissions/ folder.
2. Run python download_dependencies.py in ./kaggle_submissions/ to copy two required python package dependencies into this folder (luxai2021 and stable_baselines3).
3. Tarball the folder into a submission tar -czf submission.tar.gz -C kaggle_submissions .

Creating and viewing a replay

Place your trained model file as model.zip and your agent file agent_policy.py in the ./kaggle_submissions/ folder. Then run a command like the following from that directory:

lux-ai-2021 --seed=100 ./kaggle_submissions/main_lux-ai-2021.py ./kaggle_submissions/main_lux-ai-2021.py --maxtime 10000

This will battle your agent against itself and produce a replay match. You can view the replay here:
2021vis.lux-ai.org/