Qwen-Image InstantX Union ControlNet
Generate images with Qwen-Image InstantX ControlNet, supporting canny, soft edge, depth, pose
About
The Qwen-Image InstantX Union ControlNet workflow is designed to generate high-quality images using the Qwen-Image model with advanced ControlNet capabilities. This workflow integrates multiple control techniques such as canny edge detection, soft edge, depth, and pose estimation to offer fine-grained control over the image generation process.
By leveraging these techniques, users can create images that adhere closely to specific visual constraints or stylistic guidelines. Technically, the workflow utilizes a combination of nodes including SaveImage, LoadImage, and PreviewImage to handle image data, while specialized nodes like 0e306186-0dda-4b49-bfcb-98cab32901d9 and f8e756aa-94b4-4c02-943d-2e7910344622 are used to apply the ControlNet transformations. The FluxKontextImageScale node further refines the scaling of images, ensuring that the output maintains high fidelity to the input parameters.
What you provide
- Reading the workflow
Runs in ComfyUI. This workflow can't run on this page. Open it in Comfy Cloud with everything loaded, or download it for your own ComfyUI.
Try in CloudInside the workflow
A read-only preview of the node graph. Open it in Comfy Cloud or download it to edit the nodes.
Loading the graph
Open full-size workflow preview- Author
- ComfyUI
Requires custom nodes
- Reading the workflow
Use this workflow from code
Generated from the workflow graph and inputs when possible; stored examples are shown otherwise. Review credentials and endpoints before running code, which may incur charges.
What you need
- Runs on
- Comfy Cloud
- Key
- COMFY_API_KEY
Needs a paid Cloud plan and credits.
# Python 3.10+: pip install comfy-sdk==0.4.0
import os
from comfy_sdk import Comfy
api_key = os.environ.get("COMFY_API_KEY", "").strip()
if not api_key: raise RuntimeError("Set COMFY_API_KEY")
os.environ.setdefault("COMFY_BASE_URL", "https://cloud.comfy.org")
client = Comfy(api_key=api_key)
workflow = client.workflows.from_json({
"60": {
"class_type": "SaveImage",
"inputs": {
"filename_prefix": "ComfyUI",
"images": [
"111:8",
0
]
},
"_meta": {
"title": "Save Image"
}
},
"71": {
"class_type": "LoadImage",
"inputs": {
"image": "image_qwen_image_instantx_controlnet_input_image.jpg"
},
"_meta": {
"title": "Load Image"
}
},
"73": {
"class_type": "PreviewImage",
"inputs": {
"images": [
"112:22",
0
]
},
"_meta": {
"title": "Preview Image"
}
},
"113": {
"class_type": "FluxKontextImageScale",
"inputs": {
"image": [
"71",
0
]
},
"_meta": {
"title": "FluxKontextImageScale"
}
},
"111:3": {
"class_type": "KSampler",
"inputs": {
"cfg": 1,
"denoise": 1,
"latent_image": [
"111:76",
0
],
"model": [
"111:66",
0
],
"negative": [
"111:85",
1
],
"positive": [
"111:85",
0
],
"sampler_name": "euler",
"scheduler": "simple",
"seed": 498544006715206,
"steps": 4
},
"_meta": {
"title": "KSampler"
}
},
"111:37": {
"class_type": "UNETLoader",
"inputs": {
"unet_name": "qwen_image_fp8_e4m3fn.safetensors",
"weight_dtype": "default"
},
"_meta": {
"title": "Load Diffusion Model"
}
},
"111:38": {
"class_type": "CLIPLoader",
"inputs": {
"clip_name": "qwen_2.5_vl_7b_fp8_scaled.safetensors",
"device": "default",
"type": "qwen_image"
},
"_meta": {
"title": "Load CLIP"
}
},
"111:39": {
"class_type": "VAELoader",
"inputs": {
"vae_name": "qwen_image_vae.safetensors"
},
"_meta": {
"title": "Load VAE"
}
},
"111:6": {
"class_type": "CLIPTextEncode",
"inputs": {
"clip": [
"111:38",
0
],
"text": "Post-apocalyptic style clothing, long wavy hair, rough texture, exotic woman, tattered coarse-woven linen fabric, wearing a hood, mechanical aesthetics, mainly in dark gray tones, low-saturation earthy yellow, sense of impact and rebellion, doomsday aesthetics, grotesque aesthetics, works of art, backlighting, film photography, professional photography works, clear visible face, emotional and atmospheric dynamic photography, Fujichrome color positive film, shot with a 17mm Hasselblad ultra-wide-angle lens, f/1.2 large aperture, side backlighting, artistic light, hair light, Rembrandt light, 8K high-definition image quality, delicate real human skin texture."
},
"_meta": {
"title": "CLIP Text Encode (Positive Prompt)"
}
},
"111:66": {
"class_type": "ModelSamplingAuraFlow",
"inputs": {
"model": [
"111:80",
0
],
"sampling": "flow",
"shift": 3.1000000000000005
},
"_meta": {
"title": "ModelSamplingAuraFlow"
}
},
"111:7": {
"class_type": "CLIPTextEncode",
"inputs": {
"clip": [
"111:38",
0
],
"text": ""
},
"_meta": {
"title": "CLIP Text Encode (Negative Prompt)"
}
},
"111:76": {
"class_type": "VAEEncode",
"inputs": {
"pixels": [
"112:22",
0
],
"vae": [
"111:39",
0
]
},
"_meta": {
"title": "VAE Encode"
}
},
"111:8": {
"class_type": "VAEDecode",
"inputs": {
"samples": [
"111:3",
0
],
"vae": [
"111:39",
0
]
},
"_meta": {
"title": "VAE Decode"
}
},
"111:80": {
"class_type": "LoraLoaderModelOnly",
"inputs": {
"lora_name": "Qwen-Image-Lightning-4steps-V1.0.safetensors",
"model": [
"111:37",
0
],
"strength_model": 1
},
"_meta": {
"title": "Load LoRA"
}
},
"111:84": {
"class_type": "ControlNetLoader",
"inputs": {
"control_net_name": "Qwen-Image-InstantX-ControlNet-Union.safetensors"
},
"_meta": {
"title": "Load ControlNet Model"
}
},
"111:85": {
"class_type": "ControlNetApplyAdvanced",
"inputs": {
"control_net": [
"111:84",
0
],
"end_percent": 1,
"image": [
"112:22",
0
],
"negative": [
"111:7",
0
],
"positive": [
"111:6",
0
],
"start_percent": 0,
"strength": 1,
"vae": [
"111:39",
0
]
},
"_meta": {
"title": "Apply ControlNet"
}
},
"112:10": {
"class_type": "UNETLoader",
"inputs": {
"unet_name": "lotus-depth-d-v1-1.safetensors",
"weight_dtype": "default"
},
"_meta": {
"title": "Load Diffusion Model"
}
},
"112:14": {
"class_type": "VAELoader",
"inputs": {
"vae_name": "vae-ft-mse-840000-ema-pruned.safetensors"
},
"_meta": {
"title": "Load VAE"
}
},
"112:16": {
"class_type": "SamplerCustomAdvanced",
"inputs": {
"guider": [
"112:19",
0
],
"latent_image": [
"112:23",
0
],
"noise": [
"112:18",
0
],
"sampler": [
"112:21",
0
],
"sigmas": [
"112:28",
0
]
},
"_meta": {
"title": "SamplerCustomAdvanced"
}
},
"112:18": {
"class_type": "DisableNoise",
"inputs": {},
"_meta": {
"title": "DisableNoise"
}
},
"112:19": {
"class_type": "BasicGuider",
"inputs": {
"conditioning": [
"112:68",
0
],
"model": [
"112:10",
0
]
},
"_meta": {
"title": "Basic Guider"
}
},
"112:20": {
"class_type": "BasicScheduler",
"inputs": {
"denoise": 1,
"model": [
"112:10",
0
],
"scheduler": "normal",
"steps": 1
},
"_meta": {
"title": "BasicScheduler"
}
},
"112:21": {
"class_type": "KSamplerSelect",
"inputs": {
"sampler_name": "euler"
},
"_meta": {
"title": "KSamplerSelect"
}
},
"112:22": {
"class_type": "ImageInvert",
"inputs": {
"image": [
"112:8",
0
]
},
"_meta": {
"title": "Invert Image Colors"
}
},
"112:23": {
"class_type": "VAEEncode",
"inputs": {
"pixels": [
"113",
0
],
"vae": [
"112:14",
0
]
},
"_meta": {
"title": "VAE Encode"
}
},
"112:28": {
"class_type": "SetFirstSigma",
"inputs": {
"sigma": 999.0000000000002,
"sigmas": [
"112:20",
0
]
},
"_meta": {
"title": "SetFirstSigma"
}
},
"112:68": {
"class_type": "LotusConditioning",
"inputs": {},
"_meta": {
"title": "LotusConditioning"
}
},
"112:8": {
"class_type": "VAEDecode",
"inputs": {
"samples": [
"112:16",
0
],
"vae": [
"112:14",
0
]
},
"_meta": {
"title": "VAE Decode"
}
}
})
# run() also passes api_key to the partner nodes in this workflow.
job = client.run(workflow, api_key=api_key)Questions & answers
You need to download the diffusion model 'lotus-depth-d-v1-1.safetensors' and the VAE model 'vae-ft-mse-840000-ema-pruned.safetensors' from Huggingface or Modelscope.
This workflow is optimized for high-quality image generation rather than real-time applications. It is best suited for projects where processing time is less critical.
The choice depends on your project needs. Use canny for edge detection, soft edge for smoother transitions, depth for 3D effects, and pose for specific body positioning.
Yes, you can configure multiple nodes to apply various control techniques simultaneously, allowing for complex and customized image outputs.