Run this
pip install -r requirements.txt
** IMPORTANT ** Images download are ** NOT SAFE FOR WORK (NSFW) **
Execute this
python src/python/cli.py -h
To get a list of flags and defaults.
TL;DR
python src/python/cli.py -t <tags.txt> -dc -oc
Gets the first 100 images of each tag in the tags.txt, a newline separated file of tags to search on danbooru.
These images are saved in the directory ./gallery-dl/danbooru/<tag_name>. Then, run the face detection on each image.
This is denoted by the -d command.
Only images with one face detected are cropped and saved under the ./faces/<id>_<tag_name>.png. This is denoted by the
-c command. The -oc command is used to only crop images that have color. All cropped images are by default 64 x 64.