If you want to support me, bidding buzz on the weekly auctions page for the model helps me keep it up for on-site generation.
Recommended settings :
Positive prompt : masterpiece, best quality
Negative prompt : worst quality, low quality, bad anatomy, watermark, censored
Sampler : Euler a
CFG Scale : 5-7
I highly recommend using the ADetailer extension for better faces and eyes.
For v2, I merged v1 with some NoobAI models, @GarejiSama's AdAstra and @Manityro's Vermilion, making v2 much more compatible with NoobAI based LoRAs.
The baked in VAE is Anzhc's, which you can get here, gives a much sharper look to images, reduced noise, and makes colors less bland overall.
LustrousMix v2 generations are more colorful than v1's, with more contrast and sharpness. There is a slight art style change as well, with more mature faces.
