
This experiment was suggested by one of my Patreons over at Patreon. The idea was to see what NovelAI's image diffuser would generate if given only the name of different copyrights. An experiment I've run before among the legacy pictures, but with the new model.
I used the Danbooru API to extract all the copyright tags. This gave me a list of 34,317 copyright tags. I then filtered them to only use copyrights with 50 or more images to their name, giving us the list we have here, with some 6,963 images.
Danbooru defines copyright tags as:
Copyright tags identify the anime, manga, game, or novel that the post is associated with. More generally, copyright tags identify the source material that the characters are from.
Copyright tags should not be used for trademarks or brand names. A rule of thumb: if something doesn't have characters, it probably shouldn't be a copyright tag.
As opposed to my legacy pictures where I generated by hand, I wrote a Python script utilizing the NovelAI API to generate all these images. I first locked down a very generic prompt to allow the character tag to influence the image as much as possible and then set it up to iterate through each and every one of the tags. The script is constructed in such a way that it requires my input every time it generates a new image. So there have been a whole lot of button pressing. This is due to the NovelAI ToS, 9.1.5, which I have also confirmed with the NovelAI team.
Prompt:
1.5::______ ::, portrait, solo, -2::nude, huge breasts::, 1girl, location
Model used is NAI Diffusion V4 Full, Steps 28, Guidance 5, Quality Tags Enabled, UC Preset Heavy, Seed 3902014325.
NovelAI's V4 model is based on T5 and therefore the prompt is case sensitive1 However, all tags from Danbooru are kept in lowercase for consistency.
Since everything is generated using the very same seed, there are quite a few tags that have the potential to look good but just don’t look good here. Keep that in mind if you use these tags to generate your own prompts. Rather than looking for what actually looks good, look for a style you want and then try it out with other seeds or using your own prompt.