⇀
Return to Index

Danbooru - Artist Tags

Because someone had to test all these prompts!

Preamble

Generated example
Generated without specifying an artist in the prompt.

Since NovelAI's image generator is fine tuned using Danbooru images and tags, I wanted to see which ones it is able to reproduce. I went to the Danbooru wiki and extracted all the different artists listed.

The results was a 73 MB text file with over 280.000 artists. So I had to filter the list a bit. I went down to all artists that had at least 300 images or more to their name in a first batch (4378 artists). Then a second batch of artists with 200 to 299 images to their name (3450 artists). The end result is what you have here, 7.828 artists in total, creating a comprehensive list of artists.

Something that I notied while generating this is that V4, as compared to the original NAI Diffusion, shifts styles a whole lot even if there is an artist that it isn't trained on. That makes it very hard to filter out images that aren't relevant.

Generation

As opposed to my legacy pictures, 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 artist 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:

artist:______

Character 1

girl, soft smile, red lips, hoodie, jeans, orange hair, small breasts

Model used is NAI Diffusion V4 Full, Steps 28, Guidance 5.5, Quality Tags Enabled, UC Preset Enabled, Seed 1005398329.

Footnote

Some things to note:

Broken seeds

Due to changes in the AI model since I made the first batch, some images can't be recreated 1:1 even though the seed and settings are correct.

Weird casing on image titles!

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.

That looks weird!

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.

Jump to page: @@-Ap Ap-Ca Ca-Do Do-Ga Ga-Ha Ha-Il Im-Ka Ka-Ki Ki-La La-Me Me-Mo Mo-Ni Ni-Pa Pa-Ro Ro-Se Se-Su Su-Ti Ti-Va Va-Yu Yu-Zy Top 100