Flux Union Controlnet Pro workflow — v1.0

Use this ComfyUI workflow to guide Flux.1 D image generation with Union ControlNet Pro: supply an image for guidance, choose Depth, Pose, Canny, Tile, Blur, Grayscale, or Low quality mode, and save the generated image.

Load the workflow in ComfyUI and choose a mode in the SetUnionControlNetType node above the ControlNet loader.

For your own images, enable the OpenPose extractor by bypassing it, or replace its image with a depth or Canny extractor; select the matching mode afterward.

The archive includes the OpenPose image used for the example and the original photo from which its pose was extracted.

Update ComfyUI to its latest version and place the Union ControlNet Pro model in models/controlnet before loading the workflow.

For OpenPose extraction, the optional OpenposePreprocessor node is associated with comfyui_controlnet_aux (ControlNet Auxiliary Preprocessors).

In SetUnionControlNetType, map Canny to openpose, Tile to depth, Depth to hed/pidi/scribble/ted, and Blur to canny/lineart/anime_lineart/mlsd.

In the same node, map Pose to normal, gray to segment, and Low quality to tile.

For stronger guidance, try a strength above the example value of 0.4; the stated recommended range is 0.3–0.8, and end_percent is shown as 0.6.

Vorschlag · nicht geprüft

The workflow listing does not specify a license.