Updated: Jan 8, 2026
base model
SCROLL DOWN TO SEE VERSION DETAILS
Would appreciate comments, posts, reviews, and requests
if you'd like to pay for a commission or just send a dollar or two -> PAYPAL or Kofi
merged some loras I trained into Rectified Flow noob (loras trained on the model) - CabalResearch/NoobAI-RectifiedFlow-Experimental · Hugging Face
use the above link for recommended settings and how to get it to work, isn't just drag and drop but is near it. Comfyui workflow is in the files tab or you can click on the above link and download it there
works best with Comfyui/Swarm or Reforge
sampler: Euler A Comfy RF | or any with RF or Comfy (if using Forge, comfyui switches automatically)
schedule: simple or Normal
VAE: none or one with padding for reflection (not sure which vae to use so I mostly use none, that should use the baked vae of the RF model)
need SD3 sampling in the Advanced Model Sampling for reForge tab on reforge to be enabled. 2.5 should work.
pos: masterpiece, best quality, absurdres, highres
neg: worst quality, normal quality, bad anatomy, text,signature, artist name, watermark
Attempted to stabilize it and add a general style boost while maintaining flexibility
Version 1.A - Aesthetic
merged loras trained with experimental settings and then attempted to stabilize via a block merge with MiceandDoeV3, anatomy can be hit or miss but the colors come out clean. RF Loras work better than base noob, dunno about v-pred