First, create a folder named images and put all images inside. Prepare the resolution csv for all images, an example can be seen at resolution_information.csv. More information can be found here
├──images
├──1.jpg
├──2.jpg
├──3.jpg
├──resolution_information.csv
Then, pull the docker image and run the tool.
docker pull yukunzhou/image_automorph:latestPlease substitute the {images_path} with the path to images folder, e.g. /home/AutoMorph/images. And replace the {results_path} with the path that you want to save the results, e.g. /home/AutoMorph/Results.
docker run --rm --shm-size=2g -v {images_path}:/app/AutoMorph/images -v {results_path}:/app/AutoMorph/Results -ti --runtime=nvidia -e NVIDIA_DRIVER_CAPABILITIES=compute,utility -e NVIDIA_VISIBLE_DEVICES=all yukunzhou/image_automorph