Purpose of Module

PANNA-GVECT module demonstrates how to efficiently generate Behler-Parinello and modified Behler-Parinello descriptors (see References [1], [2], [3]).

These descriptors can then be used in machine learning algorithms. Even though these descriptors were originally designed for neural network models, they are equally suitable for other supervised learning schemes such as kernel methods, or unsupervised ones such as clustering techniques.

PANNA-GVECT, unlike other modules within the PANNA project, does not use TensorFlow framework.


PANNA-GVECT supports periodic and aperiodic structures, multiple species, derivative of the descriptors with respect to atomic positions.

Building and Testing

A stable version of the module can be downloaded using the download button on this page

As a python module PANNA-GVECT does not require installation but it relies on numpy library version >= 1.15.0.

In order to set up and test the module, run the following:

$ tar -zxvf panna-master.tar.gz
$ cd panna-master
$ python3 ./panna/


PANNA-GVECT main script requires a configuration file that specifies the parameter of the calculation such as descriptor type, length etc. A typical command for using this module is as follows:

$ export PYTHONPATH=/path/to/panna/directory/panna
$ python3 --config gvect_configuration.ini

A detailed tutorial about the contents of the configuration file can be found here.

In this comprehensive tutorial, how use this module with other modules such as PANNA-TOOLS and PANNA-TFR is also demonstrated. Together, these modules cover all the steps necessary while going from raw data to descriptors that can be used in machine learning workflow.

Source Code

PANNA-GVECT source is currently hosted on GitLab.

Further Information

The PANNA-GVECT module is developed with the contributions of R. Lot, Y. Shaidu, F. Pellegrini, E. Kucukbenli


PANNA manuscript:

[1]R. Lot, Y. Shaidu, F. Pellegrini, E. Kucukbenli. arxiv:1907.03055. Submitted (2019).


[2]J. Behler and M. Parrinello, Generalized Neural-Network Representation of High-Dimensional Potential-Energy Surfaces, Phys. Rev. Lett. 98, 146401 (2007)
[3]Justin S. Smith, Olexandr Isayev, Adrian E. Roitberg. ANI-1: An extensible neural network potential with DFT accuracy at force field computational cost. Chemical Science,(2017), DOI: 10.1039/C6SC05720A