95db34e389
* Update Japanese translation for Deeplearning4J UI (#8525) Signed-off-by: k-tamura <ktamura.biz.80@gmail.com> * RL4J: Remove processing done on observations in Policy & Async (#8471) * Removed processing from Policy.play() and fixed missing resets Signed-off-by: unknown <aboulang2002@yahoo.com> * Adjusted unit test to check if DQNs have been reset Signed-off-by: unknown <aboulang2002@yahoo.com> * Fixed a couple of problems, added and updated unit tests Signed-off-by: unknown <aboulang2002@yahoo.com> * Removed processing from AsyncThreadDiscrete Signed-off-by: unknown <aboulang2002@yahoo.com> * Fixed a few problems Signed-off-by: unknown <aboulang2002@yahoo.com> * python version bump * increase * RL4J: Replace gym-java-client with JavaCPP (#8595) * RL4J: Replace gym-java-client with JavaCPP Signed-off-by: Samuel Audet <samuel.audet@gmail.com> Co-authored-by: Kohei Tamura <ktamura.biz.80@gmail.com> Co-authored-by: Alexandre Boulanger <44292157+aboulang2002@users.noreply.github.com> Co-authored-by: Max Pumperla <max.pumperla@googlemail.com> Co-authored-by: Samuel Audet <samuel.audet@gmail.com>
81 lines
2.6 KiB
Markdown
81 lines
2.6 KiB
Markdown
# RL4J: Reinforcement Learning for Java
|
|
|
|
RL4J is a reinforcement learning framework integrated with deeplearning4j and released under an Apache 2.0 open-source license. By contributing code to this repository, you agree to make your contribution available under an Apache 2.0 license.
|
|
|
|
* DQN (Deep Q Learning with double DQN)
|
|
* Async RL (A3C, Async NStepQlearning)
|
|
|
|
Both for Low-Dimensional (array of info) and high-dimensional (pixels) input.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Here is a useful blog post I wrote to introduce you to reinforcement learning, DQN and Async RL:
|
|
|
|
|
|
[Blog post](https://rubenfiszel.github.io/posts/rl4j/2016-08-24-Reinforcement-Learning-and-DQN.html)
|
|
|
|
[Examples](https://github.com/eclipse/deeplearning4j-examples/tree/master/rl4j-examples)
|
|
|
|
[Cartpole example](https://github.com/eclipse/deeplearning4j-examples/blob/master/rl4j-examples/src/main/java/org/deeplearning4j/examples/rl4j/Cartpole.java)
|
|
|
|
# Disclaimer
|
|
|
|
This is a tech preview and distributed as is.
|
|
Comments are welcome on our gitter channel:
|
|
[gitter](https://gitter.im/deeplearning4j/deeplearning4j)
|
|
|
|
|
|
# Quickstart
|
|
|
|
* mvn install
|
|
|
|
# Visualisation
|
|
|
|
[webapp-rl4j](https://github.com/rubenfiszel/webapp-rl4j)
|
|
|
|
# Quicktry cartpole:
|
|
|
|
* run with this [main](https://github.com/eclipse/deeplearning4j-examples/blob/master/rl4j-examples/src/main/java/org/deeplearning4j/examples/rl4j/Cartpole.java)
|
|
|
|
# Doom
|
|
|
|
Doom is not ready yet but you can make it work if you feel adventurous with some additional steps:
|
|
|
|
* You will need vizdoom, compile the native lib and move it into the root of your project in a folder
|
|
* export MAVEN_OPTS=-Djava.library.path=THEFOLDEROFTHELIB
|
|
* mvn compile exec:java -Dexec.mainClass="YOURMAINCLASS"
|
|
|
|
# Malmo (Minecraft)
|
|
|
|
|
|
|
|
* Download and unzip Malmo from [here](https://github.com/Microsoft/malmo/releases)
|
|
* export MALMO_HOME=YOURMALMO_FOLDER
|
|
* export MALMO_XSD_PATH=$MALMO_HOME/Schemas
|
|
* launch malmo per [instructions](https://github.com/Microsoft/malmo#launching-minecraft-with-our-mod)
|
|
* run with this [main](https://github.com/eclipse/deeplearning4j-examples/blob/master/rl4j-examples/src/main/java/org/deeplearning4j/examples/rl4j/MalmoPixels.java)
|
|
|
|
|
|
|
|
# WIP
|
|
|
|
* Documentation
|
|
* Serialization/Deserialization (load save)
|
|
* Compression of pixels in order to store 1M state in a reasonnable amount of memory
|
|
* Async learning: A3C and nstep learning (requires some missing features from dl4j (calc and apply gradients)).
|
|
|
|
# Author
|
|
|
|
[Ruben Fiszel](http://rubenfiszel.github.io/)
|
|
|
|
# Proposed contribution area:
|
|
|
|
* Continuous control
|
|
* Policy Gradient
|
|
* Update rl4j-gym to make it compatible with pixels environments to play with Pong, Doom, etc ..
|