Qwen-Image img2img ⚡

Prompts, init latent and ControlNet attach for Qwen-Image. Leave image unconnected for txt2img. Feed the outputs straight into a stock KSampler.

By ChrisColeTech·Created 13 days ago·Updated about 3 hours ago· 6

◄prompt—►

◄negative_prompt►

◄strength0.60►

◄batch_size1►

◄width1024►

◄height1024►

◄control_modemanual►

◄depth_ckpt_namedepth_anything_v2_vitb.pth►

◄control_strength1.00►

Category🤖 CCTech/Qwen-Image

Inputs (17)

Name	Type	Default	Description
model	MODEL		—
clip	CLIP		—
vae	VAE		—
prompt	STRING		—
negative_prompt	STRING		—
strength	FLOAT	0.600–1	img2img only. How much of the init image to discard. Ignored without an image.
batch_size	INT	11–4096	—
width	INT	102416–16384	Output size. With an init or control image this resizes it.
height	INT	102416–16384	—
imageopt	IMAGE		Init image for img2img, and (in auto_canny/auto_depth modes) the source photo the control image is derived from. Leave unconnected for txt2img.
qwen_controlopt	QWEN_IMAGE_CONTROL		From QwenImageControlNetLoader.
control_modeopt	COMBO	manual	manual: connect control_image yourself. auto_canny/auto_depth: derive it from `image` automatically - pick whichever matches the loaded checkpoint. Ignored without qwen_control.
depth_ckpt_nameopt	COMBO	depth_anything_v2_vitb.pth	auto_depth mode only. Downloads on first use if not already in models/depth_anything_v2.
control_imageopt	IMAGE		Control map matching qwen_control - a canny/depth/etc map. Overrides auto_canny/auto_depth when connected. Required in manual mode.
maskopt	MASK		For an inpaint checkpoint - the region to inpaint. Not derivable automatically.
control_strengthopt	FLOAT	1.00-10–10	—
edit_referenceopt	IMAGE		For Qwen-Image-Edit checkpoints: the photo to edit. Encoded and attached to positive conditioning as the model's real edit-reference input (reference_latents) - separate from img2img's `image` and ControlNet's `control_image`. Leave unconnected for a non-Edit checkpoint.