ppo links

labml_nn/__init__.py

@@ -34,6 +34,7 @@ and
 
 #### ✨ [Sketch RNN](http://lab-ml.com/labml_nn/sketch_rnn/)
 
+#### ✨ [Reinforcement Learning](http://lab-ml.com/labml_nn/rl/)
 
 ### Installation
 

labml_nn/rl/__init__.py

@@ -0,0 +1,7 @@
+"""
+# RL Algorithms
+
+* [Proximal Policy Optimization](ppo)
+[This is an experiment](ppo/experiment.html) that runs a PPO agent on Atari Breakout.
+* [Generalized advantage estimation](ppo/gae.html)
+"""

labml_nn/rl/ppo/__init__.py

@@ -1,11 +1,8 @@
 """
-This is a an implementation of [Proximal Policy Optimization - PPO](https://arxiv.org/abs/1707.06347)
- clipped version for Atari Breakout game on OpenAI Gym.
-It runs the game environments on multiple processes to sample efficiently.
-Advantages are calculated using [Generalized Advantage Estimation](https://arxiv.org/abs/1506.02438).
+This is a an implementation of [Proximal Policy Optimization - PPO](https://arxiv.org/abs/1707.06347).
 
-*This is based on my original implementation
-[on my blog](http://blog.varunajayasiri.com/ml/ppo_pytorch.html)*.
+You can find an experiment that uses it [here](experiment.html).
+The experiment uses [Generalized Advantage Estimation](gae.html).
 """
 
 import torch

labml_nn/rl/ppo/experiment.py

@@ -1,3 +1,8 @@
+"""
+This experiment runs PPO  Atari Breakout game on OpenAI Gym.
+It runs the [game environments on multiple processes](game.html) to sample efficiently.
+"""
+
 from typing import Dict, List
 
 import numpy as np

labml_nn/rl/ppo/gae.py

@@ -1,3 +1,7 @@
+"""
+This is an implementation of paper [Generalized Advantage Estimation](https://arxiv.org/abs/1506.02438).
+"""
+
 import numpy as np