Recommended settings: Positives: <code>masterpiece, best quality,</code> Neg: <code>lowres, worst quality, low quality, bad anatomy, bad hands, 4koma, comic, greyscale, censored, jpeg artifacts</code> Dimensions: 1536x1536 / 1248x1824 (Portrait) / 1824x1248 (landscape) Sampler/scheduler: Euler A - AYS steps 27-32 CFG 6 Clipskip 2 Since we need to make it clear, these are the settings i've tested on. The resolution/dimensions are those provided by Onoma AI for Illustrious v1.0.I haven't done any extensive resolution exploration on 1.0 yet, I made something for myself so i'd be assed to train loras for 1.0 or 1.1, or whatever the model winds up settling in on and getting a finetune for.

MadlyMix V0.5 Illustrious 1.0

madlymixV05Illustrious_v05.safetensors

v0.5

A beta model featuring the aesthetics of MadlyMix NN V3 with some hacky work arounds to make it function on illustrious v1.0.Positives: masterpiece, best quality, Neg: lowres, worst quality, low quality, bad anatomy, bad hands, 4koma, comic, greyscale, censored, jpeg artifactsDimensions: 1536x1536 / 1280x1856 (Portrait) / 1856x1280 (landscape) Sampler: Euler A - Align Your Steps - feel free to experiment with other samplers, I just haven't gotten much testing done outside of these. Steps: 27-32 CFG 6 Clipskip 2

MadlyMix V0.5 - Illustrious XL 1.0

Positives: masterpiece, best quality,

Neg: lowres, worst quality, low quality, bad anatomy, bad hands, 4koma, comic, greyscale, censored, jpeg artifacts

Dimensions: 1536x1536 / 1280x1856 (Portrait) / 1856x1280 (landscape)
Sampler/Scheduler: Euler A - AYS
Steps: 27-32 
CFG 6
Clipskip 2 

Files with SHA256: 078df7dd70b8f3235fb730beabf1c64f78bbf12633a06b0475003943e68f4d45