Have a ComfyUI workflow? Get a readable Workflow Doc in your browser.Make a workflow readable.Try it now ↗
I'm sharing this slightly modified WF with this tutorial from the author. It's a bit slow, but it gives me good results, better than SD Upscale. You can, of course, use the original WF from the comfy UI (explained in...
Runtime profile
Source description
❌Steudio Upscale
I'm sharing this slightly modified WF with this tutorial from the author. It's a bit slow, but it gives me good results, better than SD Upscale. You can, of course, use the original WF from the comfy UI (explained in the video).
Estimated VRAM requirement
Estimate unavailable
11.1 GB across 2 of 8 model files. Model file total + 25% loading overhead + 2 GB execution buffer, rounded up.
Requirements
15 requirementsAI-generated commentary
AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.
Steudio Upscale NDDG — v2.0 is a workflow for upscaling an input image and saving the resulting image.
The workflow’s graph includes image loading, encoding, sampling, decoding, tile handling, ControlNet, preview, comparison, and saving.
LoadImage supplies the image input; SaveImage supplies the saved image output; PreviewImage and Image Comparer (rgthree) support preview and comparison.
Version v2.0 is published by NeoDroleDeGueule, and Flux.1 D is the listed base model.
Before use, check that these model files are available: fluxmania_Legacy.safetensors; FLUX_shakker_labs_union_Controlnet.safetensors; t5xxl_fp8_e4m3fn_scaled.safetensors; and 4x-UltraSharpV2.pth.
Suggestion · not verified
Also check these model entries: flux_vae.safetensors; clipLCLIPGFullFP32_zer0intVisionCLIPL.safetensors; microsoft/Florence-2-base; and fluxDevFp8AnalogArt_v10VersionCheckpoint.safetensors.
Suggestion · not verified
Check that these node packs are available: kaytool, jameswalker-nodes, teacache, comfyui-florence2, comfyui_steudio, comfyui-post-processing-nodes, and rgthree-comfy.
Suggestion · not verified
The graph includes tile division and combination, ControlNet application, color adjustment, film grain, and image comparison.
It also includes bypass and mute controls, a Power Lora Loader, and TeaCache.
The publisher describes the workflow as “a bit slow”; treat that as reported context rather than a guaranteed runtime.
Suggestion · not verified
No license is listed for this workflow.
Does this need to be edited?
Sign in to send an edit request.
Sources
1 sourceSource excerpts
2 excerptsSource context: 244 downloads · Type Workflows · Base model Flux.1 D
Upscaler · Hugging Face · GeraldoZulimar/4x-UltraSharpV2
clipLCLIPGFullFP32_zer0intVisionCLIPL.safetensors
Not resolvedText encoder · Unknown
FLUX_shakker_labs_union_Controlnet.safetensors
Not resolvedControlNet · Unknown
VAE · Hugging Face
fluxDevFp8AnalogArt_v10VersionCheckpoint.safetensors
Not resolvedCheckpoint · Unknown
Checkpoint · 6.30 GB · SAFETENSOR · fp16 · Hugging Face · civitai2026/flux1-dev-checkpoints
Vision model · Hugging Face · microsoft/Florence-2-base
Text encoder · 4.80 GB · Hugging Face · comfyanonymous/flux_text_encoders
Node pack · Registry
comfyui-florence2
PossibleNode pack · Registry
Source context: 239 downloads · Type Workflows · Base model Flux.1 D
Node pack · Registry
jameswalker-nodes
Not resolvedNode pack · Registry
Node pack · Registry
Node pack · Registry
teacache
Not resolvedNode pack · Registry
