Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
This workflow is designed to help beginners easily upscale existing images using the FLUX model with the Ultimate SD Upscale node.
Runtime profile
Source description
This workflow is designed to help beginners easily upscale existing images using the FLUX model with the Ultimate SD Upscale node.
It’s a very simple setup, stripped of any additional features like LoRA or complex functions, focusing solely on upscaling and image comparison.
This makes it easy to understand and perfect for those new to the process.
Additionally, I’ve included notes within the workflow specifying the required models for each node. Be sure to check them out for smoother usage!
This workflow requires the Image Compare node and the Ultimate SD Upscale node to function properly. If these nodes are not already installed, you can easily add them using the Install Missing Node feature in Comfyui.
AI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
FLUX+Ultimate SD Upscale (Basic Workflow) is for upscaling existing images with FLUX and Ultimate SD Upscale; it takes an existing image and produces an upscaled image, with image comparison included.
The simple setup focuses on upscaling and image comparison without LoRA or complex functions.
Image handling covers loading, comparing, previewing, and saving through LoadImage, Image Comparer (rgthree), PreviewImage, and SaveImage.
Input: an existing image. Output: an upscaled image.
The workflow includes image comparison, preview, and saving through Image Comparer (rgthree), PreviewImage, and SaveImage.
Image Compare and Ultimate SD Upscale are required; if either is missing, the instructions point to ComfyUI’s Install Missing Node feature.
Place the FLUX model in C:\ComfyUI_windows_portable\ComfyUI\models\unet\FLUX and the VAE file in C:\ComfyUI_windows_portable\ComfyUI\models\vae.
The named model files are flux1-dev-Q8_0.gguf, t5xxl_fp16.safetensors, clip_l.safetensors, FLUX_VAE_diffusion_pytorch_model.safetensors, flux1DevFp8_v10.safetensors, and 4xFaceUpSharpDAT.pth.
The named nodes include CLIPTextEncodeFlux, DualCLIPLoader, Image Comparer (rgthree), LoadImage, PreviewImage, and SaveImage.
They also include UltimateSDUpscale, UNETLoader, UnetLoaderGGUF, UpscaleModelLoader, and VAELoader.
Place text encoder files in C:\ComfyUI_windows_portable\ComfyUI\models\clip and the downloaded .pth upscale model in C:\ComfyUI_windows_portable\ComfyUI\models\upscale_models.
Use the notes inside the workflow to check which model each node requires.
Suggestion · not verified
Do not use lllyasviel/flux1-dev-bnb-nf4; the instructions say it cannot be used.
Suggestion · not verified
Before using it, verify that Image Compare, Ultimate SD Upscale, and the needed model files are available.
Suggestion · not verified
Does this need to be edited?
Sign in to send an edit request.
Sources
1 sourceSource excerpts
1 excerptSource context: 1452 downloads · Type Workflows · Base model Flux.1 D
[Required Nodes]
Image Compare
Ultimate SD Upscale
[Model Link]
[Direct Links to Download Models]
Dev GGUF Download
Schnell GGUF Download
Note: lllyasviel/flux1-dev-bnb-nf4 cannot be used.
(Model Directory)
Place the model in:
C:\ComfyUI_windows_portable\ComfyUI\models\unet\FLUX
Download from:
(Directory for VAE)
Place the file in:
C:\ComfyUI_windows_portable\ComfyUI\models\vae
This model is used in the DualclipLoader node.
(Directory for Text Encoders)
Place the file in:
C:\ComfyUI_windows_portable\ComfyUI\models\clip
After downloading the .pth file, place it in the following directory:
C:\ComfyUI_windows_portable\ComfyUI\models\upscale_models
============================================================
FLUX 모델을 활용해 기존 이미지를 Ultimate SD Upscale 노드로
간단히 업스케일할 수 있도록 설계되었습니다.
로라(LoRA)나 복잡한 기능은 모두 제외하고,
업스케일과 이미지 비교에만 집중한 매우 간단한 구성이라 초보자도 쉽게 이해하고 활용할 수 있습니다.
또한, 각 노드에 필요한 모델 정보를 메모에 작성해 두었으니
참고하시면 더욱 편리하게 사용하실 수 있습니다!
이 워크플로우를 작동하려면 Image Compare 노드와 Ultimate SD Upscale 노드가 추가로 필요합니다.
만약 이 노드들이 설치되어 있지 않다면, Install Missing Node 기능을 사용해 간단히 설치할 수 있습니다.
[필요한 노드]
Image Compare
Ultimate SD Upscale
[모델 다운로드 다이렉트 링크 모음]
Dev GGUF 다운로드
Schnell GGUF 다운로드
참고: lllyasviel/flux1-dev-bnb-nf4는 사용할 수 없습니다.
(모델 저장 경로)
모델 파일을 아래 경로에 넣으세요:
C:\ComfyUI_windows_portable\ComfyUI\models\unet\FLUX
다운로드 링크:
(VAE 저장 경로)
파일을 아래 경로에 넣으세요:
C:\ComfyUI_windows_portable\ComfyUI\models\vae
이 파일은 DualclipLoader 노드에서 사용됩니다.
(텍스트 인코더 저장 경로)
파일을 아래 경로에 넣으세요:
C:\ComfyUI_windows_portable\ComfyUI\models\clip
.pth 파일을 다운로드한 후, 아래 경로에 넣으세요:
C:\ComfyUI_windows_portable\ComfyUI\models\upscale_models
[contact]
Youtube : https://www.youtube.com/@A01demort
Instagram : https://www.instagram.com/a01demort/
Naver Cafe : https://cafe.naver.com/sdfkorea
Estimated VRAM requirement
Estimate unavailable
21.2 GB across 3 of 10 model files. Model file total + 25% loading overhead + 2 GB execution buffer, rounded up.
Requirements
17 requirements4xFaceUpSharpDAT.pth
Not resolvedUpscaler · Unknown
Text encoder · 235 MB · SAFETENSORS · Unknown
DualCLIPLoader
Not resolvedText encoder · Unknown
FLUX_VAE_diffusion_pytorch_model.safetensors
Not resolvedVAE · Unknown
Checkpoint · 11.8 GB · GGUF · fp16 · Unknown
flux1DevFp8_v10.safetensors
Not resolvedCheckpoint · Unknown
Text encoder · 9.12 GB · SAFETENSORS · Unknown
UltimateSDUpscale
Not resolvedUpscaler · Unknown
UpscaleModelLoader
Not resolvedUpscaler · Unknown
VAELoader
Not resolvedVAE · Unknown
CLIPTextEncodeFlux
Not resolvedNode pack · Unknown
Image Comparer (rgthree)
Not resolvedNode pack · Unknown
LoadImage
Not resolvedNode pack · Unknown
Node pack · Unknown
SaveImage
Not resolvedNode pack · Unknown
UNETLoader
Not resolvedNode pack · Unknown
UnetLoaderGGUF
Not resolvedNode pack · Unknown
