Commit 44e12f6d authored by Julian Johannes Steinsberger-Dührßen's avatar Julian Johannes Steinsberger-Dührßen
Browse files

changes to installation script for mxnet

parent 38974c6a
Pipeline #327136 passed with stage
in 6 minutes and 18 seconds
......@@ -9,19 +9,23 @@ pip3 install --user --upgrade "cmake>=3.13.2"
wget https://bootstrap.pypa.io/get-pip.py
python get-pip.py
pip install --user h5py matplotlib numpy==1.16.5 mxnet==1.5.1 #Not alll test work with mxnet v1.6.0, the curent standard version installed for the cpu installation of mxnet.
#You could alternativly also use python 3.6 instead of 2.7, then you could also use the newest numpy version.
#Note that you then have to also set the PYTHON_PATH acordingly, or specifiy the python path for all applications in their build scripts
#and test currently only run on the python specified in the PYTHON_PATH.
#If you want to use mxnet with cuda install f.e. mxnet-cu100 for cuda 10.0 (for this currently v1.5.1 is already the newest version),
#of course then you have to install cuda and cudnn beforehand.
pip install --user h5py matplotlib numpy==1.16.5 mxnet==1.5.1.post0 #The newest version installed the curent standard version installed v1.7.0 cant be run with python2 the current standard of the EMDAL2CPP generator,
#As the needed numpy version is not suported anymore for python2 (python will no longer be supported).
#Further more not all test work with mxnet v1.6.0. And when just using v1.5.1 you can't compile against the libmxnet.so needed for compiling
#the cpp prediction part, the same holds for 1.6.0 but not 1.5.1.post0 and versions later than 1.6.0 (1.7.0) as there was some compression done for this
#library which was then droped again by the developers of mxnet.
#You could alternativly also use python 3.6 instead of 2.7, then you could also use the newest numpy version.
#Note that you then have to also set the PYTHON_PATH acordingly, or specifiy the python path for all applications in their build scripts
#and test currently only run on the python specified in the PYTHON_PATH.
#If you want to use mxnet with cuda install f.e. mxnet-cu100 for cuda 10.0 (if v1.5.1 is the newest version (here it specifing post0 is not
#neccescery), otherwise no gurantee for the numpy dependency, see above), #of course then you have to install cuda and cudnn beforehand.
git clone --recursive https://github.com/apache/incubator-mxnet.git mxnet
cd mxnet && git checkout tags/1.5.0 && git submodule update --recursive --init
cd mxnet && mkdir build && cd build && cmake -DUSE_CPP_PACKAGE=1 -DUSE_CUDA=0 -GNinja .. && ninja -v
cd mxnet && cp -r include/mxnet /usr/include/mxnet && cp -r cpp-package/include/mxnet-cpp /usr/include/ && cp -r 3rdparty/tvm/nnvm/include/nnvm /usr/include/ && cp -r 3rdparty/dmlc-core/include/dmlc /usr/include/
#you have tohave armadillo-9.600.6.zip in your current folder
#you have to have armadillo-9.600.6.zip in your current folder
unzip armadillo.zip -d .
cd armadillo-9.600.6 && cmake . && make && make install
......
......@@ -7,7 +7,7 @@ RUN pip3 install --user --upgrade "cmake>=3.13.2"
RUN wget https://bootstrap.pypa.io/get-pip.py
RUN python get-pip.py
RUN pip install --user h5py matplotlib numpy==1.16.5 mxnet==1.5.1
RUN pip install --user h5py matplotlib numpy==1.16.5 mxnet==1.5.1.post0
RUN git clone --recursive https://github.com/apache/incubator-mxnet.git mxnet
RUN cd mxnet && git checkout tags/1.5.0 && git submodule update --recursive --init
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment