Commit 8f23f359 authored by Emine Kucukbenli's avatar Emine Kucukbenli

updated the PANNA-TRAIN module, minor edit in other PANNA modules

parent 9b9b1459
Pipeline #4780 passed with stages
in 1 minute and 49 seconds
......@@ -66,7 +66,7 @@ A detailed tutorial about the contents of the configuration file can be found
`here <https://gitlab.com/PANNAdevs/panna/blob/master/doc/tutorial/README_tutorial_2_data_preparation.md>`_.
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 whlie going from raw data to descriptors that can be
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
......
......@@ -65,7 +65,7 @@ A detailed tutorial about the contents of the configuration file can be found
`here <https://gitlab.com/PANNAdevs/panna/blob/master/doc/tutorial/README_tutorial_2_data_preparation.md>`_.
In this comprehensive tutorial, how use this module with other modules such as PANNA-GVECT and PANNA-TOOLS
is also demonstrated. Together, these modules cover all the steps necessary whlie going from raw data to descriptors that can be
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
......
......@@ -43,23 +43,57 @@ It further supports controlling the training dynamics: eg. reeze/unfreeze layers
Building and Testing
______________________________
A stable version of the module can be downloaded using::
XXX
Current installation and testing are done with::
XXX
A stable version of the module can be downloaded using the download button on this `page <https://gitlab.com/PANNAdevs/panna>`_
Here are the commands for installation::
XXX
As a python module PANNA-TRAIN does not require installation but it relies on numpy library version >= 1.15.0, tensorflow version >= 1.13.0, and
tensorboard version >= 1.13.0. Note that with version 2.0.0, tensorflow libraries went under substantial changes in structure, the 1.1X.X
family supports the equally valid previous structure and is still being maintained. PANNA-TRAIN requires tensorflow 1.1X.X family of versions.
Test::
XXX
In order to set up and test the module, run the following::
$ tar -zxvf panna-master.tar.gz
$ cd panna-master
$ python3 ./panna/test-train.py
Usage
______
PANNA-TRAIN main script requires a configuration file that specifies the parameter of the calculation
such as number of layers and nodes of each neural network layer, learning parameter etc.
A typical command for using this module is as follows::
$ export PYTHONPATH=/path/to/panna/directory/panna
$ python3 train.py --config train_configuration.ini
A detailed tutorial about the contents of the configuration file can be found
`here <https://gitlab.com/PANNAdevs/panna/blob/master/doc/tutorial/README_tutorial_1_training.md>`_.
In this comprehensive tutorial, a neural network training scenarios is demonstrated from beginning to end.
Network validation is a key step in network training, hence in the tutorial how to use this module together
with PANNA-EVAL module used in validation is also explained.
Together, these two modules cover all the steps necessary to train an atomistic neural network, starting from a data which specifies
the machine learning task in (input, target output) paird form.
Source Code
___________
To be released soon.
PANNA-TRAIN source is currently hosted on `gitlab <https://gitlab.com/PANNAdevs/panna>`_.
Further Information
______________________
The PANNA-TRAIN module is developed with the contributions of R. Lot, Y. Shaidu, F. Pellegrini, E. Kucukbenli
References
____________
PANNA manuscript:
[1] R. Lot, Y. Shaidu, F. Pellegrini, E. Kucukbenli.
`arxiv:1907.03055 <https://arxiv.org/abs/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
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment