Atomvision: A deep learning framework for atomistic image data
First create a conda environment: Install miniconda environment from https://conda.io/miniconda.html Based on your system requirements, you'll get a file something like 'Miniconda3-latest-XYZ'.
Now,
bash Miniconda3-latest-Linux-x86_64.sh (for linux)
bash Miniconda3-latest-MacOSX-x86_64.sh (for Mac)
Download 32/64 bit python 3.6 miniconda exe and install (for windows) Now, let's make a conda environment, say "version", choose other name as you like::
conda create --name version python=3.8
source activate version
Now, let's install the package:
git clone https://github.com/usnistgov/atomvision.git
cd atomvision
pip install torchvision
python setup.py develop
This example shows how to classify 2D-lattice (5 classes) for 2D-materials STM/STEM images.
Datasets can be generated with STM/STEM sections of the data folder with generate_data.py script or pre-populated image datasets can be downloaded with 'download.py`. We create two folders train_folder, ``test_folder`` with sub-folders ``0,1,2,3,4,...`` for individual classes and they contain images for these classes in way train-test splits have proportionate amount of images.
An example for using pre-trained densenet on STEM JARVIS-DFT 2D dataset is given below. Change ``train_folder`` and ``test_folder`` paths in order to use a different dataset.
python atomvision/scripts/train_classifiers.py --model_name densenet --train_folder atomvision/data/classification/stem_jv2d/train_folder --test_folder atomvision/data/classification/stem_jv2d/test_folder
Note: the repository is under development.
Please cite the following if you happen to use JARVIS-Tools for a publication.
Choudhary, K. et al. The joint automated repository for various integrated simulations (JARVIS) for data-driven materials design. npj Computational Materials, 6(1), 1-13 (2020).