Hilda (Style LoRA) | NoobAI-XL V-Pred 1.0

No trigger word required.

Use your usual generation settings, but if you want what I use it's this:

• Gradient Estimation Sampler

• KL Optimal Scheduler

• 50-100 steps

• 3.5-5 CFG

• 0.1-0.2 SAG

• 3 PAG

• If doing Hires-fix; Lanczos upscale + ~0.6 de-noise