Microstructure Inpainter

Microstructure inpainter is a python app for inpainting material science microstructural images using GANs. The app can be run via a command line interface or a graphical interface. Although the GANs can be trained on CPU, GPU training is advised.

Quickstart

To run from executable file. Download the file specific for your operating system.

MacOS users - unzip, and then double click the icon and allow access to folders.

Linux users - unzip and then run from command line

./path/to/inpainter

To run locally.

Prerequisites:

• conda
• python3

Create a new conda environment, activate and install pytorch

Note: cudatoolkit version and pytorch install depends on system, see PyTorch install for more info.

conda create --name inpainter
conda activate inpainter
conda install pytorch torchvision -c pytorch
pip install -r requirements.txt

If you are planning to use Weights and Biases to track training runs create a .env file to hold secrets, the .env file must include

WANDB_API_KEY=
WANDB_ENTITY=
WANDB_PROJECT=

You are now ready to run the repo. To run the GUI

python run_gui.py

Alternatively, to start training a model using the CLI, use the following command

python run_cli.py train -t <name> -c <x1 x2 y1 y2> -p <path/to/data> -i <image_type> -s <method> -w <WANDB>

-t takes the name of the training run, -c takes the coordinate values as a vector of 4 numbers representing the corners of a rectangle, -p takes the path to the data, -i takes the image type (n-phase, grayscale, colour), -s takes the method name (rect, poly), -w takes a boolean whether to track run with Weights and Biases. For example,

python run_cli.py train -t test -c <360 440 360 440> -p <data/example_inpainting.png> -i n-phase -s rect -w False

To generate samples from a trained generator

python run_cli.py generate -t test -c <360 440 360 440> -p <data/example_inpainting.png> -i n-phase -s rect -w False

Saving, loading and overwriting models

Models are saved to runs folder which is generated when training initiates. Inside runs, a new folder with the name of your training run tag will be generated, inside this the model params and training outputs are saved. This includes:

• config.json - this json holds the config paramaters of your training run, see config.py for more info
• Gen.pt - this holds the generator training parameters
• Disc.pt - this holds the discriminator parameters

Training

If training for the first time, these files are created and updated during training.

If you initiate a training run with a tag of a run that already exists you will see the prompt

To overwrite existing model enter 'o', to load existing model enter 'l' or to cancel enter 'c'.

By entering 'o' you will overwrite the existing models, deleting their saved parameters and config. l will load the existing model params and config, and continue training this model. c will abort the training run.

Evaluation

When evaluating a trained model, the params and model config are loaded from files. Models are saved with their training tag, use this tag to evaluate specific models.

Roadmap

• Windows single exe
• Save and load trained models
• Implement both methods for both shapes