Bogdan Mazoure

ppo_jax. Jax implementation of Proximal Policy Optimization (PPO) specifically tuned for Procgen, with benchmarked results and saved model weights on all environments.

dreamerv2_jax. Dreamerv2's re-implementation (and slight architectural improvement) in Jax + Flax

ctrl_public. Code for "Cross-Trajectory Representation Learning for Zero-Shot Generalization in RL"

sparseMuJoCo. Sparse environment for MuJoCo suite (v2 and v3)

DRIML. Code for Deep Reinforcement and InfoMax Learning (Neurips 2020)

pixel_world. Python

gsf_public. Preliminary code for Representation learning with Generalized Similarity Functions

ms_pacman_gym. Partially observable PacMan game in OpenAI Gym format

batch_stationary_distribution. Implementation of "Batch stationary distribution estimation" via variational power method

code_snippets. Jupyter Notebook