Nodes/ComfyUI-FatMex-Nodes/Fat Mex Image Edit Sampler
ComfyUI Node

Fat Mex Image Edit Sampler

All-in-one Qwen Image Edit sampler. Feed reference images + a prompt describing the edit. Handles vision-language encoding, layered latent creation, sampling, and decoding in one node.

By FatMex·Created 6 months ago·Updated 6 months ago· 3
Fat Mex Image Edit Sampler
  • model
  • clip
  • vae
  • image1
  • image2
  • image3
  • images
  • latent
prompt
negative_promptugly, blurry, deformed, bad anatomy, low quality
seed0
preset_hintauto
steps0
cfg0.0
sampler_nameauto
schedulerauto
width1024
height1024
layers3
CategoryFat Mex

Inputs (17)

NameTypeDefaultDescription
modelMODELModel from Fat Mex Model Loader (Qwen Edit preset).
clipCLIPCLIP from Fat Mex Model Loader.
vaeVAEVAE from Fat Mex Model Loader.
promptSTRINGDescribe the edit. Reference images as 'image 1', 'image 2', etc.
negative_promptSTRINGugly, blurry, deformed, bad anatomy, low qualityWhat to avoid.
seedINT00–18446744073709550000
image1optIMAGEReference image 1 (e.g. the face to use).
image2optIMAGEReference image 2 (optional).
image3optIMAGEReference image 3 (optional).
preset_hintoptCOMBOauto8 options: auto, Klein 9B, Klein 9B True, Qwen Image Edit 2509, Qwen Image Edit 2511, Qwen Image 2512, +2
stepsoptINT00–100Override steps (0 = preset).
cfgoptFLOAT0.00–100
sampler_nameoptCOMBOauto45 options: auto, euler, euler_cfg_pp, euler_ancestral, euler_ancestral_cfg_pp, heun, +39
scheduleroptCOMBOauto10 options: auto, simple, sgm_uniform, karras, exponential, ddim_uniform, +4
widthoptINT102464–16384Output width.
heightoptINT102464–16384Output height.
layersoptINT30–16Qwen layered latent layers (usually = number of images).

Outputs (2)

NameTypeDescription
imagesIMAGE
latentLATENT