Img2Img-ControlNetXL-Union — Img2Img AI Model
One ControlNet XL model covering six conditioning modes — canny, depth, openpose, hed, normal and segment — so a control image plus a prompt steers SDXL without swapping checkpoints. Outputs up to 1024×1024. The control image can be up to 1024×1024.
What it does
ControlNet XL Union is a single adapter covering six conditioning modes - canny, depth, openpose, hed, normal and segment - so a control image plus a prompt steers SDXL without swapping a checkpoint for each type.
Pick the mode that matches the control image you have: an edge map holds the outlines, a depth map holds the layout, a pose skeleton holds the figure. The prompt supplies everything else, from style to lighting to material. Output goes up to 1024x1024.
When to use it
- Keeping a photograph's composition while changing its style entirely
- Placing a character in a chosen pose, from an openpose skeleton
- Restyling a scene while its 3D layout stays put, guided by depth
- Locking what goes where with a segmentation mask, prompt fills in how
- Line art from a canny or HED map turned into a rendered image
Details
- CategoryImg2Img
- Year2024
- LicenseApache-2.0
Compliance & Provenance
- ProviderOpen-source (xinsir) · Specialized
- EU AI Act RiskMinimal Risk
- Art. 50 TransparencyNot applicable
Inputs & Outputs
- Control ImageInputimage
Preprocessed control image matching the selected condition mode: a canny edge map, depth map, pose skeleton, HED edge map, normal map, or segmentation mask. Use an Annotator node to generate this from a source image. The control image can be up to 1024×1024.
- TextInputstring
Text prompt describing the desired output image. SDXL responds best to detailed English prompts. Include subject, style, lighting, and mood. Negative prompt is available as an advanced parameter.
- ImageOutputimage
Generated image guided by the control image and text prompt, at the specified width × height.
Parameters
- condition_modestringdefault cannyoptions: canny, depth, openpose, hed, normal, segment
Conditioning type matching the control image (canny/depth/openpose/hed/normal/segment).
- controlnet_conditioning_scalefloatdefault 0.8range 0–2
Control guidance strength (0 = ignore, 2 = strict). 0.7–1.0 typical.
- stepsintegerdefault 25range 20–30
Diffusion steps (20–30).
- guidance_scalefloatdefault 5range 1–20
CFG scale (1–20, 5.0–7.5 typical for SDXL).
- seedintegerdefault -1range -1–2147483647
Random seed (-1 = random).
- widthintegerdefault 1024options: 512, 768, 1024
Output width (px). Control image resized to match.
- heightintegerdefault 1024options: 512, 768, 1024
Output height (px). Control image resized to match.
- negative_promptstringdefault advanced
Negative prompt — what to avoid. Empty for none.
Tags
Alternatives in Image Control
- ControlNet XL Canny
Guided SDXL image generation from a Canny edge map + text prompt. Sharper structural adherence than Union. Outputs up to 1024×1024.
- Depth Anything Annotator
Monocular depth estimation (Depth Anything V1 Large).