Model-free Reinforcement Learning in Tensorflow 2

Algorithms

Soft Actor-Critic
Twin-delayed DDPG
Proximal Policy Optimization

Features

One file per algorithm so that they can serve as submodules for other projects without dependencies
Results for MuJoCo and PyBullet

Requirement

Training

Visualization

Reference