Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
Simple workflow with ControlNet tiled upscale
Runtime profile
Source description
Simple workflow with ControlNet tiled upscale
This is my main workflow for SDXL/Pony/Illustrious that uses ControlNet tile upscaling instead of Ultimate SD. Feel free to DM me if you have any questions about the workflow.
v1.2 update
Added CFG Skimming, after more testing I think it's an interesting tool. Can be easily disabled in generation settings options (enable_skimming). Change CFG to recommended checkpoint values when you disable it.
Saved images should now be compatible with Civitai. I was getting tired of filling some fields manually. Requires manual installation of Save Images With Metadata Universal Extension
Combined generation settings into a single subgraph
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Foton's Workflow with ControlNet upscale is for SDXL, Pony, or Illustrious image generation with ControlNet tiled upscaling instead of Ultimate SD; it takes generation and upscale settings and can produce saved images with Civitai-compatible metadata when Save Images With Meta-
It uses four node packs: ComfyUI-Impact-Subpack, Skimmed_CFG, ComfyUI-Impact-Pack, and ComfyUI-TiledDiffusion.
The settings subgraph includes controls for the initial upscale model, ControlNet tile checkpoint, tile batch size, blend factor, upscale size, and denoise.
Face detailing accepts custom positive and negative prompts, and saved images can include metadata for Civitai compatibility when Save Images With Metadata Universal Extension is installed manually.
Install Save Images With Metadata Universal Extension manually, then run Rules Scanner Subgraph once after installation or after changing nodes in ComfyUI so the metadata saver works properly.
Place the ControlNet models in ComfyUI/models/controlnet.
The workflow lists these node packs: comfyui-impact-pack, ComfyUI-TiledDiffusion, comfyui-impact-subpack, and skimmed_cfg.
Before use, check for these exact files: GOBOLightingILL.safetensors, glinty_v1.safetensors, TTPLANET_Controlnet_Tile_realistic_v2_fp16.safetensors, remacri_original.safetensors, and bbox/face_yolov8m.pt.
Suggestion · not verified
Verify that the node packs are available under the exact names ComfyMath and SaveImageWithMetaDataUniversal.
Suggestion · not verified
To disable CFG Skimming, turn off enable_skimming and use the checkpoint's recommended CFG values.
Suggestion · not verified
For 3–4x upscaling, use denoise 0.2 or lower; for lower VRAM, use Tiled Diffusion batch size 1 and VAE Decode tile size 512. Batch size 1 reduces VRAM load at the cost of speed.
Suggestion · not verified
For face detailing, set Use custom prompt to true to enter positive and negative prompts; increase feather if visible edges appear around faces.
Suggestion · not verified
No license is listed for this workflow, so verify the applicable usage terms before using it.
Suggestion
Changing overlap in the 512×512 low-VRAM setup is explicitly untested.
Does this need to be edited?
Sign in to send an edit request.
Sources
1 sourceSource excerpts
1 excerptSource context: 348 downloads · Type Workflows · Base model Illustrious
For MetaData saver to work properly, run Rules Scanner Subgraph once after installation or when you change nodes in ComfyUI (I don't make the rules)
By default set tile_batch_size to 1 in HiRes subgraph to safely run on https://huggingface.co/TTPlanet/TTPLanet_SDXL_Controlnet_Tile_Realistic
Alternative model for anime and stylized models - sharper and cleaner details: https://huggingface.co/xinsir/controlnet-union-sdxl-1.0
Put these models into ComfyUI/models/controlnet folder
KSampler can be replaced with a ComfyUI one, I'm using Inspire's node for different noise generation options.
Control upscale size and denoise in the HiRes subgraph
For higher upscale values (3-4x) use lower denoise (0.2 or even lower).
Custom Adetailer prompt
In Use custom prompt node set value to true and type positive/negative prompts which will be used for faces. Userful when some Loras overpower Adetailer and change face color too much, or if you just want to adjust face expression. With additional nodes you can set it up to use different Loras during this step. Increase feather value if you get visible edges around faces.
Low VRAM
To save VRAM inside a HiRes Upscale subgraph try this:
In Tiled Diffustion node Set batch size to 1. This will reduce VRAM load considerably at the cost of speed.
Set VAE Decode tile size to 512.
If it's still VRAM hungry, try setting tile width/height in Tiled Diffusion and VAE encode to 512x512. Might want to change overlap as well, but I haven't tested this.
Estimated VRAM requirement
Estimate unavailable
2.39 GB across 2 of 5 model files. Model file total + 25% loading overhead + 2 GB execution buffer, rounded up.
Requirements
11 requirementsObject Detector · Hugging Face · bbox/face_yolov8m.pt
glinty_v1.safetensors
Not resolvedCheckpoint · Unknown
GOBOLightingILL.safetensors
Not resolvedLoRA · Unknown
Upscaler · 63.8 MB · SAFETENSORS · Unknown
ControlNet · 2.33 GB · SAFETENSORS · fp16 · Unknown
ComfyMath
PossibleNode pack · Registry
Node pack · Registry
Node pack · Registry
Node pack · Registry
SaveImageWithMetaDataUniversal
Not resolvedNode pack · Registry
Node pack · Registry
