Changelog 18/4/25:
Second version, way more consistent and with better prompt adherence from my testing. I reworked the entire dataset aswell as the settings used to train it, and I think I've hit a sweet spot, so hopefully it'll have better results for y'all.
Recommended
Strength: 0.9, lower strength for more flexibity. As for the cfg, from my testing using something around 4.5, even 5.0 works fine, but you can always lower it for a more painterly effect aswell.
Thank you again for using my LoRA, you can support me through either donating Buzz, or you can use one of the links provided beneath my profile.