RL

Reinforcement learning algorithms, built with nbml

Contents

• ppo.rs Continuous Proximal Policy Optimization
• td3.rs Twin Deep Delayed Deterministic Policy Gradients
• discrete_ppo.rs Gumbel Softmax Discrete Proximal Policy Optimization
• sac.rs Soft Actor-Critic (WIP)
• ddac3.rs Advantage Actor Critic with autoregressive state/action prediction through a generative transformer decoder policy (Custom)