We construct a 3D graph neural network model XAS3D to simulate XANES. It turns to be faster than the traditional XANES fitting method when we combine the simulation model and XANES optimization algorithm to fit the 3D structure of the given system. IHEP BSRF Zhaohf Group

conda create -n pyg_pl python==3.9
source /scratchfs/heps/zhanf/miniconda3/bin/activate pyg_pl
pip install pytorch-lightning==1.8 -i https://mirrors.ustc.edu.cn/pypi/web/simple
pip install tqdm

pip install torch==1.13.1+cu117 torchvision==0.14.1+cu117 torchaudio==0.13.1+cu117 -f https://download.pytorch.org/whl/torch_stable.html -i https://pypi.tuna.tsinghua.edu.cn/simple
pip install torch-sparse torch-scatter -f https://pytorch-geometric.com/whl/torch-1.13.1+cu117.html
pip install torch-geometric -f https://pytorch-geometric.com/whl/torch-1.13.1+cu117.html
pip install torch-cluster torch-spline-conv -f https://pytorch-geometric.com/whl/torch-1.13.1+cu117.html
pip install torchmetrics==0.7

pip install sympy
pip install nlopt
pip install pymatgen
pip install frechetdist

example_Fe3O4: python Fit.py
example_Fe_element: python Fit.py
example_Mndoped: python Fit.py
First, copy the dataset from the "datasets" folder to the "example" folder, or change the dataset file location in the code.
Then run "python Fit.py"

example_Fe3O4/Fe3O4_XAS3Dabs_hyper : python BAT_gpu_hyper.py
Run "python BAT_gpu_hyper.py" to batch submit the running scripts to the Slurm cluster.