Setup for a project/competition amongst students to train a winning Reinforcement Learning agent for the classic game Bomberman.
Branch for the main agent (master)
The agent had to be zipped so that it could be uploaded to github, since github does not allow upload of files larger than 100 MByte. Therefore, the agent has to be unzipped into the my_q_agent directory, before running.
Branch for the second agent (second_architecture)
During development, we created many branches which by now are deprecated. The only two branches used are, as already described, the branch "master", for main agent as well as the branch "second_architecture" for the second architecture. Deployment of the main agent is tested in Docker with the provided Docker file.