muhangtian / rl-bandits

RL & Bandits

Project Strucutre

• All runs are initiated with main.py
• algo contains my own implementation of some algorithms like DQN and PPO
• model folder contains the neural network architectures used in the learning algorithms (which are stored in algo)
• trained contains trained agents (neural networks) stored in .pt format using PyTorch, as well as video recordings of agent behavior
• utils contains environment wrappers, validation functions, data loaders, and preprocessing functions, which are used in algorithms implemented in algo
• cluster contains shell scripts used to run experiments using cluster computer (with slurm)