This code was used to generate the figures in the article
D. Neumann, T. Wiese, and W. Utschick, Learning the MMSE Channel Estimator, Accepted for publication in IEEE Transactions on Signal Processing, 2018.
-
Download Julia version 0.6.4 from https://julialang-s3.julialang.org/bin/linux/x64/0.6/julia-0.6.4-linux-x86_64.tar.gz (note- there has been substancial syntax changes to julia 1.0+, this repo will not work with julia 1.0+ or julia 0.7. However, if one need to migrate the code to latest julia version, one must fix all the depricated warning given by juila 0.7)
-
Unizip:
tar -xvzf julia-0.6.4-linux-x86_64.tar.gz
Add to path: export PATH="$PATH:/path/to/<Julia directory>/bin"
, give path to the bin folder in julia directory.
- Installing packages-
Pkg.add("Packagename")
install the following packages: "CSV", "DataFrames", "Interpolations", "Distributions" .
- Tensorflow.jl installation:
4.1 Frist check if python 3.5 is installed on the system and accessible to the directory. Install tensorflow version 1.4.1 on this using
pip3 install tensorflow==1.4.1
and donnot update this tensflow package.
4.2 Now install tensorflow.jl which is a wrapper for tensorflow package for python. It uses pycall to call the tf functions.
Pkg.add("TensorFlow",v"0.8.0",v"0.9.0")
This is essential for installing TensorFlow.jl version 0.8.0, as it will install >=0.8.0 and <0.9.0, which is just equivalent to saying install version= 0.8.0. It is essential that one installs TensorFlow.jl version 0.8.0 as higher version will lead to to segmentation fault in this repository.
Can check the packages installed and their respective versions using: Pkg.status()
removing package: Pkg.rm("package name")
for more information of tensorflow.jl refer: https://github.com/malmaud/TensorFlow.jl
Julia installation and all dependencies are there in the julia directory itself, so for removing julia from the system just do rm -rf <julia directory>
Now remove the folder containg this repo from the system