Interesting. This is a big out of my field lol.

I appreciate you helping out with the community engagement side of things , Allo.

What I’d really like is for people to use my methods on their own generators.

That way I can use what they have created and not have to worry about debugging it myself.

Would be nice to have some more user engagement though.

I think people still too confused about my generator to ask questions lol .

Yesterday I sent links to chats of the “main” ai-text-to-image generator and furry-ai.

The furries were cool , which is funny in contrast absolute mind-boggling depravity I found in the e621 dataset. I want those people on my genarator.

Meanwhile , individuals on the main “ai-text-to-image” seemed to be absolutely nuts. I don’t want those. No way!

So I think “quality” vs “quantity” matters too :)