Flux ControlNet Upscaling Workflow with Florence2 and GGUF Supported

Welcome back, everyone! In this video, I’m breaking down an advanced ComfyUI upscaling workflow using the Flux model—perfect for anyone looking to dramatically enhance image quality, even on low-memory systems.

🔗Models:

🔗 Nodes:

[jdesanti]

After many tests, I realized that the seed in the KSampler was fixed, changing it to random solved my problems

[Dany-wg]

Wow, you're a genius! Thanks, bro!

[BuckeyeGuy]

FYI: This workflow is only viable for upscales up to something less than 2k. If you go higher than that with this workflow, you're going to go OOM.

[wonder]

I also have a RTX 3090, and to avoid initial out of memory error, I dropped the factor of upscale from 7 to 4. This allowed case sampler to start running. 15 minutes later and it’s only finished by 25%. The fans on my computer are going wild and this for an upscale that’s only 4K, how could this possibly be worth the time, or the expense in electricity? You really need to fix this workflow. Thanks anyway.

[phenix]

Thank work great for under 2k picture (even if it take time for 2k with my 3080 10go.)
but could you work on a version of the worflow working for low memory card (like 10go, in my case) to be able to output 4k picture (like 3840px width or height depending if it's a portrait or landscape) without getting oom it would be awesome ?
because with the current workflow anythink above 2k, li something close to the 3840px in W or H doesn't even launch the ksampler process ( i mean seeing any % ) i'm using the ggulf due to the limit of my gc, even after 15 min of waiting .

[kryz]

Nice tutorial but where do i need to put this file: Florence-2-pixelprose, give me all the path please.

[MilindNarvekar]

VAEEncode
convolution_overrideable not implemented. You are likely triggering this with tensor backend other than CPU/CUDA/MKLDNN, if this is intended, please use TORCH_LIBRARY_IMPL to override this function

how to fix this?

[laurent]

Getting "No module named 'peft'" error on the DownloadAndLoadFlorence2Model node.
What am I doing wrong ? (edit: I installed peft, but I'm using portable SD so it's not finding it. Do I need to install in the portable folder ?)

[xyzxyz]

i have RTX3090 (24 VRAM) and 128GB system RAM, i still have error of "torch.cuda.OutOfMemoryError: Allocation on device 0"..

[AlanBurns-oc]

Neat, thanks. I can see some uses for this, but it's really unsuitable for humans if you want them to look remotely like themselves. SUPIR does humans much better but isn't good with very low resolutions. This is good with low resolutions but bad with humans. But if you just want generic people in the image, this is fantastic.

[neararsan]

Error occurred when executing ControlNetLoader:

MMDiT.__init__() got an unexpected keyword argument 'image_model'

please help

[taezonday]

Cant load the VAE in the workflow. How to unhide the selection? Just says load VAE with no options.

[sasansasani]

i have NVIDIA GeForce RTX 4070 Ti, Total VRAM 12282 MB, total RAM 65277 MB, i still have error of Allocation on device 🥲🥲🥲 why???

[mr.entezaee]

Why is it so slow now? What is death? Our father is getting old to take this picture with quality! is it normal

[Djonsing]

This design requires 50GB of video memory.