Just download the imported package in the .py and .ipynb.
Download the traning dataset locally, and run the script in dataProcess.ipynb. (Remember to modify your path of data, which in noticed by TODO in dataProcess.ipynb)
cd model
python train.pyRemember to modify your path of data, which in noticed by TODO in train.ipynb.
| Model \ Dataset | Water Splitting (eV) |
|---|---|
| CGCNN | 1.623 |
| CrystalGraphBert | 3.80 |
PS:
The original testing data of Water Splitting for CGCNN can't be found because of the change of the dataset. The data is from https://cmr.fysik.dtu.dk/cubic_perovskites/cubic_perovskites.html .
This project builds upon the work presented in the following paper:
Paper: Zhang J, Zhang H, Xia C, et al. Graph-bert: Only attention is needed for learning graph representations[J]. arXiv preprint arXiv:2001.05140, 2020.
Code: https://github.com/jwzhanggy/Graph-Bert
Data: Dataset Water Splitting
XiYuan Project in Fudan University.
Advisor: Prof. Yiqiang Zhan, Prof. Xia Cai
Student: Zhijian Zhou, Yang Shen
