https://github.com/divelab/AIRS

conda create --name comformer python=3.10
conda activate comformer
conda install pytorch torchvision torchaudio pytorch-cuda=11.6 -c pytorch -c nvidia # or higher version if you want
conda install pyg -c pyg
pip install jarvis-tools==2022.9.16
python setup.py



训练

cd comformer/scripts
python train_jarvis.py # for jarvis
python train_mp.py # for the materials project

https://github.com/divelab/AIRS

环境设置
我们将用于安装依赖项和设置环境。我们建议使用 Python 3.9 Miniconda 安装程序。conda
安装完成后，运行conda
conda env create -f environment.yml
然后运行以下代码以激活环境
conda activate potnet
我们正在用于实现我们的图神经网络。安装 via conda 似乎有问题，因此我们明确使用 pip 作为torch-geometrictorch-geometric
pip install torch-scatter2.1.0 torch-sparse0.6.16 torch-cluster1.6.0 torch-spline-conv1.2.1 -f https://data.pyg.org/whl/torch-1.12.1%2Bcu116.html
和

pip install dgl1.1.1
pip install pydantic1.10.2
pip install cython3.0.0
pip install ase3.22.1
pip install jarvis-tools2022.9.16
pip install numpy1.21.6
pip install pandarallel1.6.3
pip install pandas1.3.5
pip install periodictable1.6.1
pip install pymatgen2021.3.3
pip install pytorch-ignite0.4.10
pip install pyyaml6.0
pip install scikit-learn1.0.2
pip install scipy1.7.3
pip install tqdm==4.64.1

pip install torch-geometric
请注意，我们正在使用旧版本的 JARVIS 工具包。最新的 JARVIS 工具包将包含新版本的数据集，其中包含的数据比我们在论文中介绍的数据更多。jarvis-tools==2022.9.16
运行求和算法
要运行求和算法，请运行以下命令以安装算法包（请记住将其替换为您自己的目标。TARGET_PATH
cd functions
tar xzvf gsl-latest.tar.gz
cd gsl-2.7.1
./configure --prefix=TARGET_PATH
sudo apt install libgsl-dev
make
make install
Then edit by adding~/.bashrc
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:TARGET_PATH/lib/
和

source ~/.bashrc
or create another terminal.

Now we back to directory and runfunctions
python setup.py build_ext --inplace
Then the algorithm is installed as Cython package. A simple way to test if it is successfully installed is to run below in the root directory.
python test_algorithm.py