Personal machine learning projects to tinker with various ML technologies. More folders will be created/updated as more projects are added.
This program uses Reinforced learning and HPO to train an AI to progress as far as possible through the original super mario bros game. Utilizes Pytorch, Tensorflow, Stable_Baselines3, nes_py, gym, and Optuna.
Currently a work in progress. This program utilizes Reinforced Learning and Curriculm Learning to train an AI the importance of different aspects of the game Doom. It utilizes VizDoom (https://github.com/mwydmuch/ViZDoom) for the rendering of the game as well as its configuration files that are used to train different learning curriculms. Currently only suited to perform in basic configurations. Will be improved to work with more advanced configurations as well as the full Doom game. Utilizes Pytorch, Tensorflow, Stable_Baselines3, gym, and vizdoom.