Torch can easily be installed from the repository and includes LUAJit:
git clone https://github.com/torch/distro.git ~/torch-master --recursive cd ~/torch-master bash install-deps # Installation will take longer with CUDA and CuDNN: ./install.sh
At the end, the installer will add the following line to the bash profile:
After starting a new console, the installation can be tested by running the interactive Torch console:
th ______ __ | Torch7 /_ __/__ ________/ / | Scientific computing for Lua. / / / _ \/ __/ __/ _ \ | Type ? for help /_/ \___/_/ \__/_//_/ | https://github.com/torch | http://torch.ch th> require 'torch' # will list all the functions ...
# Console should be restarted after installing Torch: sudo apt-get install libzmq3-dev libssl-dev python-zmq sudo apt-get install luarocks luarocks install lzmq --local luarocks install image --local git clone https://github.com/facebook/iTorch.git iTorch-master cd iTorch-master luarocks make --local
The kernel can be tested by running
jupyter notebook and looking for the iTorch kernel.
ZeroBrane Studio can easily be installed from the website or as follows:
wget https://download.zerobrane.com/ZeroBraneStudioEduPack-1.50-linux.sh chmod +x ZeroBraneStudioEduPack-1.50-linux.sh sudo ./ZeroBraneStudioEduPack-1.50-linux.sh rm ZeroBraneStudioEduPack-1.50-linux.sh
In order to run Torch from within ZeroBrane Studio, it needs to know where Torch's binary is located:
By starting ZeroBrane Studio from console (after creating a new console instance) using
zbstudio, Torch can easily be used; test it using