sprakashdash / rl.fun.do

A repository for easy understanding of codes in Deep Reinforcement Learning

RL.Fun.Do

A repository for easy understanding of codes in Deep Reinforcement Learning

This tutorial presents latest research in policy gradient methods for modle-free RL in the following order:

1. Advantage Actor Critic (A2C)

• A3C Paper
• OpenAI blog

2. Continuous control with deep reinforcement learning

• DDPG Paper

3. Proximal Policy Optimization Algorithms

• PPO Paper
• OpenAI blog

4. Trust Region Policy Optimization

• OpenAI Spinning Up
• TRPO Paper

5. Soft Actor Critic Algorithm

• BAIR UC Berkeley Blog
• SAC 2018 Paper
• SAC 2019 Paper

I have also updated Soft Actor Critic with tensorboard feature to plot your learning curves.

Here is one of the training sample for MuJoCo HalfCheetah:

If you want a hang in PyTorch then refer my tutorials to get started or the official tutorials.

Best RL courses to refer:

• Berkeley Deep RL Course (2019)
• Deep RL Bootcamp
• David Silver's course
• Practical RL