Skip to content

Instantly share code, notes, and snippets.

@anis016
Last active March 8, 2018 13:08
Show Gist options
  • Save anis016/0f9c759e57dd641cd18e67516475dbed to your computer and use it in GitHub Desktop.
Save anis016/0f9c759e57dd641cd18e67516475dbed to your computer and use it in GitHub Desktop.
# building rl4j for the latest snapshot (mvnrepository still doesn't have the latest build)
# build pipeline >> javacpp -> libnd4j -> nd4j -> DataVec -> deeplearning4j -> rl4j
# https://github.com/deeplearning4j/rl4j/issues/73#issuecomment-335732797
0. note about maven version. it must be latest (running 3.5.0)
1. install cmake
1.1 wget https://cmake.org/files/v3.11/cmake-3.11.0-rc2-Linux-x86_64.sh
1.2 chmod a+x cmake-3.6.2-Linux-x86_64.sh
1.3 mkdir /opt/cmake
1.4 sudo sh cmake-3.11.0-rc2-Linux-x86_64.sh --prefix=/opt/cmake --skip-license
1.5 ln -s /opt/cmake/bin/cmake /usr/local/bin/cmake
1.6 cmake --version
2. clone the repos
2.1 git clone https://github.com/bytedeco/javacpp.git
2.2 git clone https://github.com/deeplearning4j/libnd4j.git
2.3 git clone https://github.com/deeplearning4j/nd4j.git
2.4 git clone https://github.com/deeplearning4j/DataVec.git
2.5 git clone https://github.com/deeplearning4j/deeplearning4j.git
2.6 git clone https://github.com/deeplearning4j/rl4j.git
3. export LIBND4J_HOME=/home/user/directory/libnd4j # this is the path of the libnd4j in bashscript
4. build the repos serially (no cuda)
4.1 cd javacpp, mvn clean install -DskipTests
4.2 cd libnd4j, mvn clean install -DskipTests
4.3 cd nd4j, mvn clean install -DskipTests -pl '!nd4j-backends/nd4j-backend-impls/nd4j-cuda,!nd4j-backends/nd4j-backend-impls/nd4j-cuda-platform'
4.4 cd DataVec, mvn clean install -DskipTests
4.5 cd deeplearning4j, mvn clean install -DskipTests -pl '!./deeplearning4j-cuda/'
4.6 cd rl4j, mvn clean install -DskipTests
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment