agat 7.13.4.2

Atomic Graph ATtention networks for predicting atomic energies and forces.

AGAT (Atomic Graph ATtention networks)

Using AGAT

The documentation of AGAT API is available.

Installation

Install with conda environment

• Create a new environment

  conda create -n agat python==3.10
  

• Activate the environment

  conda activate agat
  

• Install package

  pip install agat
  

• Install dgl.
  Please navigate to the Get Started page of dgl. For GPU version:

  conda install -c dglteam/label/cu118 dgl
  

  For now, the cpu version 1.1.2 of dgl has bugs. You can install the cpu version with pip install dgl==1.1.1.

• Change dgl backend to tensorflow.

  If you still cannot use tensorflow backend dgl, run the following on Linux OS:

  wget https://data.dgl.ai/wheels/cu118/dgl-1.1.1%2Bcu118-cp310-cp310-manylinux1_x86_64.whl
  pip install ./dgl-1.1.1+cu118-cp310-cp310-manylinux1_x86_64.whl
  pip install numpy --upgrade
  

• For tensorflow of GPU version, if you don't have CUDA and CUDNN on your device, you need to run (Linux OS):

  conda install -c conda-forge cudatoolkit=11.8.0
  pip install nvidia-cudnn-cu11==8.6.0.163
  mkdir -p $CONDA_PREFIX/etc/conda/activate.d
  echo 'CUDNN_PATH=$(dirname $(python -c "import nvidia.cudnn;print(nvidia.cudnn.__file__)"))' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
  echo 'export LD_LIBRARY_PATH=$CUDNN_PATH/lib:$CONDA_PREFIX/lib/:$LD_LIBRARY_PATH' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
  source $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
  # Verify install:
  python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"
  

  Refer to Install TensorFlow with pip and Tensorflow_GPU for more details (other OSs).

Quick start

Prepare VASP calculations

Collect paths of VASP calculations

• We provided examples of VASP outputs at VASP_calculations_example.
• Find all directories containing OUTCAR file:

  find . -name OUTCAR > paths.log
  

• Remove the string 'OUTCAR' in paths.log.

  sed -i 's/OUTCAR$//g' paths.log
  

• Specify the absolute paths in paths.log.

  sed -i "s#^.#${PWD}#g" paths.log
  

Build database

from agat.data import AgatDatabase
if __name__ == '__main__':
    ad = AgatDatabase(mode_of_NN='ase_dist', num_of_cores=2)
    ad.build()

Train AGAT model

from agat.model import Train
at = Train()
at.fit_energy_model()
at.fit_force_model()

Model prediction

from agat.app import GatApp
energy_model_save_dir = os.path.join('out_file', 'energy_ckpt')
force_model_save_dir = os.path.join('out_file', 'force_ckpt')
graph_build_scheme_dir = 'dataset'
app = GatApp(energy_model_save_dir, force_model_save_dir, graph_build_scheme_dir)
graph, info = app.get_graph('POSCAR')
energy = app.get_energy(graph)
forces = app.get_forces(graph)

Geometry optimization

from ase.io import read
from ase.optimize import BFGS
from agat.app import GatAseCalculator
from agat.default_parameters import default_hp_config
poscar = read('POSCAR')
calculator=GatAseCalculator(energy_model_save_dir,
                            force_model_save_dir,
                            graph_build_scheme_dir)
poscar = Atoms(poscar, calculator=calculator)
dyn = BFGS(poscar, trajectory='test.traj')
dyn.run(**default_hp_config['opt_config'])

Some default parameters

agat/default_parameters.py; Explanations: docs/sphinx/source/Default parameters.md.

Change log

Please check Change_log.md