Skip to content


TensorX is written in pure python but it's made to complement Tensorflow, so it depends on Tensorflow, which needs to be installed from the tensorflow package. The reason for this is that you might want to install Tensorflow builds optimized for your machine (see these). Additionally, TensorX has optional dependencies like matplotlib or pygraphviz for certain functionality.

with pip

Install using pip with the following commands

pip install tensorflow 
pip install tensorx 

You can install a custom tensorflow wheel using it's URL like so:

pip install

with Poetry

TensorX is easy to install using Poetry which manages the dependencies and a virtualenv for you. See Poetry documentation for details on installation and usage. Once Poetry is installed, you can use it to install TensorX either from PyPI, or directly from git.

If you want to create a new project using poetry, simply run:

poetry new myproject
cd myproject

if you already have a project, move to it's directory and run the following commands:

poetry add tensorflow
poetry add tensorx
# or, if you want to have tensorx as a dependency from the git repository
poetry add git+ 

Developer Installation

For a developer installation of TensorX, simply clone the git repository and install it with Poetry. The git repository has a pyproject.toml and poetry.lock files. These allow for the installation to be reproducible --meaning that Poetry will install the exact versions of dependencies that were being used on a specific commit.

# Clone TensorX git repository
git clone
cd tensorx

# Install dependencies 
poetry install

# You need Tensorflow which is not specified as a dependency
# to install it without adding it as a dependency simple run
poetry run pip install tensorflow

poetry run will run a given command inside the project current virtualenv