Why are you lying (results are from novelai-pruned)
I'm trying to wrap my head around how many times you've lied and posted this model.
Firstly, it's perfectly fine to train your own model trained on DDLC or whatever gelbooru/r34 sites you scrape (even when danbooru2021 exists). That's not my problem, nor the virus detection flare-ups from pickled files. I dont use reddit and I'm currently at 1 karma in order to enter the DDLCr34 club and uncover a lump of shit.
I don't get why you'd need to lie about your model results. You're showing off "incremental improvements" and getting some reddit karma on the results, and only if users "use the universal negative prompt" you could somehow make an infinite monika dreambooth/model of your own. (https://www.reddit.com/r/StableDiffusion/comments/y6yajr/comment/isvzq9f/)
...if the results were legitimate
First image you'd see of this "model"; run it though PNG info in the 1111 webUI, and it shows that it's made from the leaked NovelAI weights:
I was gonna go on with my life and not care that someone lies about how good their model is until I peeked at your replies after proudly taunting "you get karma on reddit" in the GitHub. Holy shit, was I impressed. Six fucking years on that platform and people are hating your guts. 10s of Monika images posted the same day, credited as "source: me" and "source: self-made AI", both of which are verifiable lies, and to which you're constantly called out for it.
Your spam is pissing normal people off and fuelling the social media shitstorm of AI hate.
(https://www.reddit.com/r/DDLCRule34/comments/y6r5fv/comment/isqu3y2/)
Please, for the sake of any sort of creative effort that's making you want to produce as much Monika art as possible. Tell the truth. Seriously, make a model, post AI art (when appropriately tagged), share results; just don't lie about it ffs. Fucking 6 years on reddit, and you're still trying to farm karma.