VoidMix

An SDXL diffusion model merged and tuned from Illustrious-XL, crafted specifically for atmospheric, volumetric, and high-fidelity stylized anime illustration.

NOTICE

CivitAI weirdly lists the required model files, go to the official huggingface page and download the files from there, and put them in the proper folders.

Generation settings:

• Sampler: euler_ancestral

• Steps: 25-30 (28 is best)

• CFG Scale: 5.0-7.0

• Clip skip: 2 (Set last layer to -2)

• VAE: sdxl.vae.safetensors

• CLIP: Dual Clip (illustrious_clip_l & illustrious_clip_g)