INSANIUM v1
Crafted through precision block-layer merging LoRAs from personally curated datasets in order to invoke surprising fidelity, style versatility, and prompt adherence.
Built on top of the versatile powerhouse; Illustrious 2.0, this model excels when generating at higher resolutions. 1.25 - 1.5 megapixels will remove your need for detailers, and create deep, intricate scenes.
For best results, Use this model with my Iterative Noise KSampler Node! It uses a unique method of noise injection per a user-determined step interval which enables Illustrious models, and this model in particular, to create far more fine details with impressive coherence.
my preferred settings:
5.5 cfg
40 steps
Sampler: Euler Ancestral
Scheduler: sgm_uniform
add noise per steps: 2
starting noise: 0.24
ending noise: 0.00
1280 x 1280 total pixels