An AI-powered package for polymer informatics
- Polymer generator
- generate random polymers that can be used for unsupervised learning purposes
- as a Prior, which can be optimized by algorithms like reinforcement learning later, for polymer design purposes
PolymerAI is written in pure Python, but it does not support installation from pip or conda at current stage. More convenient ways for installation will be built in the future.
At current stage, please download the package directly from github or use
$ git clone https://github.com/RUIMINMA1996/PolymerAI.gitto get the package.
PolymerAI consists of several modules, and in order to run each module, please go to the module's folder and excute files following the instruction in that folder.
run "python agent_baseline_steptest.py" to generate polymers with high thermal conductivity.
pretrained model can be downloaded via the following link: https://drive.google.com/file/d/1gAwUq4T5MxFXnRSqmANtvAgTv7slcoc8/view?usp=sharing
To cite this repository:
@software{PolymerAI2020github,
author = {Ruimin Ma, and Tengfei Luo},
title = {{P}olymerAI: An AI-Powered Package for Polymer Informatics},
url = {https://github.com/RUIMINMA1996/PolymerAI},
version = {1.0},
year = {2020},
}