DeepMosaics VS doujinci

Instructions are here and in the README of the repository. https://imgur.com/a/D1EQYkE https://gitlab.com/ubcseagull/doujinci

TL;DR Provide a nhentai link or id, wait 15-20 minutes, 80-90% of the bars or mosaics will be decensored automatically for free. Remember HentAI and DeepCreamPy? If not, the gist was that DeepCreamPy was this great machine learning tool that could decensor bars and mosaics as long as you colored them in green with Gimp or Photoshop. HentAI came along to do this coloring automatically (also with machine learning), which was great, except that you still had to download the doujin images yourself, convert them to PNG, remove screentones, and then use two clunky UIs to use HentAI and DCP (not to mention fiddle with Python dependencies sometimes). I saw an opportunity to use modern CI (Continuous Integration) tools like GitLab and Docker to do all the menial work automatically, and thus I made DoujinCI. In short, I stripped away all the frontend of HentAI and DCP, added the automatic downloading, jpg-png conversion, screentone removal, and configured it to work in a GitLab pipeline. Now, pretty much everyone can drastically decensor any doujin for free. Instructions are here and in the README of the repository. https://imgur.com/a/D1EQYkE https://gitlab.com/ubcseagull/doujinci Enjoy! Note: you're limited to 400 free pipeline minutes per month by GitLab, but you can set up your own runner on your PC if you want to bypass that. If you're decensoring more than 20 doujins a month though, it might be time to touch some grass.