download model and install tutorial: https://comfyanonymous.github.io/ComfyUI_examples/z_image/
Uncensor node: https://github.com/nova452/ComfyUI-ConditioningKrea2Rebalance
enter: git clone https://github.com/nova452/ComfyUI-ConditioningKrea2Rebalance
inside if your custom node folder if comfy manager dosen't have it yet.
Description
2.0 had some bugs and missing links and the donutnode had a bug that broke editing when i uploaded it. and i made this so it edits be default for those who want to try that out with proper upscaling and face detailer. grounding_px controls likeness vs frexibility ("grounding resolution (default 768; trained range 512–1536). This is a quality dial: lower = stronger edit adherence, higher = stronger identity/likeness. Try 1024+ for people, 512 for stubborn scene changes.")
FAQ
Comments (8)
.... please help me... im stuck cause i cant find the loras
umm question... i cant find some of the loras and the donut stuff doesnt work for me.... probably the wrong vesion... can you include more details/update your site for more information
thx
loras will dn automatically if the nodes work haha
@DonutsDelivery idk they dont :/ and i only get like colorful static (denoise) at the end
QUESTION: where do we link the "reference" node ? I see it near the "Edit" node, but while the Edit one is connected, the reference is not. Also how does it work? And then I see on top a node called: "1 = off, 2 = on, what is it related to?
the reference is the original krea2 edit stuff. it dosent make use of the new edit lora, and its not really an "edit" at all but something that will try to force through concepts from a reference image. its got each their toggles underneath the input image. not sure if they work together tbh as edit mode uses its own conditioning path. what you prob want is to have the workflow support multiple images for the edit lora. that is coming. was a bit pressed on time and wanted to get this out so people could try it. ive got some other stuff to fix for the noise scheduling aswell.
got prompt
[ERROR] Failed to validate prompt for output 63:
[ERROR] CLIPLoader 430:
[ERROR] - Value not in list: type: 'krea2' not in ['stable_diffusion', 'stable_cascade', 'sd3', 'stable_audio', 'm
ochi', 'ltxv', 'pixart', 'cosmos', 'lumina2', 'wan', 'hidream', 'chroma', 'ace', 'omnigen2', 'qwen_image', 'hunyuan_
image', 'flux2', 'ovis', 'longcat_image', 'cogvideox']
[ERROR] SeedGenerator 195:
[ERROR] - Failed to convert an input value to a INT value: fixed_seed, randomize, invalid literal for int() with b
ase 10: 'randomize'
[ERROR] - Value not in list: mode: 214897638556888 not in ['Random', 'Fixed']
[ERROR] Output will be ignored
[ERROR] Failed to validate prompt for output 43:
[ERROR] DonutTiledUpscale 664:171:
[ERROR] - Return type mismatch between linked nodes: seed, received_type(STRING) mismatch input_type(INT)
[ERROR] DonutFaceDetailer 664:255:
[ERROR] - Return type mismatch between linked nodes: seed, received_type(STRING) mismatch input_type(INT)
[ERROR] Output will be ignored
[ERROR] Failed to validate prompt for output 71:
[ERROR] Output will be ignored
[ERROR] Failed to validate prompt for output 124:
[ERROR] Output will be ignored
[ERROR] Failed to validate prompt for output 64:
[ERROR] Output will be ignored