XenonPy project

Build Status Build Status Build status codecov Version Python Versions Downloads PyPI - Downloads

XenonPy is a Python library that implements a comprehensive set of machine learning tools for materials informatics. Its functionalities partially depend on PyTorch and R. The current release provides some limited modules:

XenonPy inspired by matminer: https://hackingmaterials.github.io/matminer/.

XenonPy is a open source project https://github.com/yoshida-lab/XenonPy.

See our documents for details: http://xenonpy.readthedocs.io


  1. Yamada, H. et al. Predicting Materials Properties with Little Data Using Shotgun Transfer Learning. ACS Cent. Sci. acscentsci.9b00804 (2019). doi:10.1021/acscentsci.9b00804

XenonPy images

XenonPy images packed a lot of useful packages for materials informatics using. The following table list some core packages in XenonPy images.

Package Version
PyTorch 1.4.0
tensorly 0.4.4
pymatgen 2020.3.13
matminer 0.6.2
mordred 1.2.0
scipy 1.4.1
scikit-learn 0.22.2
xgboost 1.0.0
ngboost master
fastcluster 1.1.26
pandas 1.0.3
rdkit 2019.03.3
jupyter 1.0.0
seaborn 0.10.0
matplotlib 3.2.1
plotly 4.5.4


In order to use this image you must have Docker Engine installed. Instructions for setting up Docker Engine are available on the Docker website.

CUDA requirements

If you have a CUDA-compatible NVIDIA graphics card, you can use a CUDA-enabled version of the PyTorch image to enable hardware acceleration. This only can be used in Ubuntu Linux.

Firstly, ensure that you install the appropriate NVIDIA drivers and libraries. If you are running Ubuntu, you can install proprietary NVIDIA drivers from the PPA and CUDA from the NVIDIA website.

You will also need to install nvidia-docker2 to enable GPU device access within Docker containers. This can be found at NVIDIA/nvidia-docker.


Pre-built xenonpy images are available on Docker Hub under the name yoshidalab/xenonpy. For example, you can pull the CUDA 10.0 version with:

docker pull yoshidalab/xenonpy:cuda10

The table below lists software versions for each of the currently supported Docker image tags .

Image tag CUDA PyTorch
latest 10.1 1.4.0
cpu None 1.4.0
cuda10 10.1 1.4.0
cuda9 9.2 1.4.0

Running XenonPy

It is possible to run XenonPy inside a container. Using xenonpy with jupyter is very easy, you could run it with the following command:

docker run --rm -it \
  --runtime=nvidia \
  --ipc=host \
  --volume=$Home/.xenonpy:/home/user/.xenonpy \
  --volume=<path/to/your/workspace>:/workspace \

Here's a description of the Docker command-line options shown above:

You may wish to consider using Docker Compose to make running containers with many options easier. At the time of writing, only version 2.3 of Docker Compose configuration files supports the runtime option.

Copyright and license

┬ęCopyright 2020 The XenonPy project, all rights reserved. Released under the BSD-3 license.