Semilustrious

CHECKPOINT
Original


Updated:

An attempt at pulling Illustrious towards semi realism. Worked well enough for my use case, I use it for original character design/concept ideation.

Works best with hires.

Versions are not 'updates' they're different styles.

  • V1 - More anime and 2D with clear line art and shading.

  • V2 - More realistic/ semi realistic than anime. you can push it either way.

Add to prompt: masterpiece, best quality, amazing quality,

(works without these as well)

Ratings: general, sensitive, explicit

(no guarantees of course)

Works well with DMD2 lora + LCM beta sampler for faster results.

4 steps + hires with 5-6 steps. (CFG at 1, Hires denoise at 0.3 - 3.5)

If you don't like DMD2 or need to squeeze more creativity and prompt adherence,

12-15 steps + hires with 15-25 steps with Euler or DPMPP2. (CFG 5 - 8, Hires denoise 0.3 - 0.65)

Resolutions that I use:

  • 512x768 hires 2x

  • 680x1024 hires 1.5x - 2x

  • 512x1024 hires 15x - 2x

  • 384x1024 hires 1.5x - 2x

Version Detail

Illustrious
more semi realistic illustrious. add to prompt: masterpiece, best quality, amazing quality, ratings: general, sensitive, explicit etc Works well with DMD2, LCM beta sampler. 512x768 hires 2x. Or 680x1024 hires 1.5x. (or higher)

Project Permissions

    Use Permissions

  • Use in TENSOR Online

  • As a online training base model on TENSOR

  • Use without crediting me

  • Share merges of this model

  • Use different permissions on merges

    Commercial Use

  • Sell generated contents

  • Use on generation services

  • Sell this model or merges

Related Posts