Hello there and thanks for checking out this workflow!
—Intro—
SUPIR is a very powerful detailing/image-restoration model, capable of producing mind-blowing results, but usually requires some tweaking to do so,...
unless you have a setup that just works™, no matter the input.
Use-cases :
• (Hyper-)detailing generated images, even if they already are of high resolution and quality
• Photo-restoration tasks to get low-res and blurry content up to scratch again.
—Features—
Installation and download guide for models and nodes
►INFO guides explaining various settings
Upscaling
Florence 2 + WD14 prompting assistance
100 Steps to SUPIR — Journey to become a better image
—Custom Nodes used—
All of which can be installed through the ComfyUI-Manager
—Troubleshooting—
If nodes show up red (failing to load), check the 'Install Missing Custom Nodes' tab of the ComfyUI Manager for the missing node packs and install them.
Please check if all custom node packs load properly after installing, i.e. no
(IMPORT FAILED)messages next to any of them in the console upon ComfyUI startup.Always reload/drag'n'drop the original, downloaded workflow file into ComfyUI to reload an intact version of the workflow.
→ The last opened workflow that appears on startup shows a cached version of the workflow, "remembering" group nodes that failed due to missing nodes as failed, keeping them broken even after having everything installed correctly.
—Thanks—
The workflow would not be possible as is without these custom node packs. If you want to support the custom node creators, give them a ⭐ on their github repos! Thank you!
Feel free to ask any questions, share improvements or suggestions in the comment section!
Also let me know if you encounter any confusing points I can elaborate on and focus on improving for the next update!
Description
v2
— replaced bypass control with mute
— replaced image output size display
— addition of image input options
— addition of Florence2 captioning
— custom node packs changes :
• addition : ComfyUI-Easy-Use, ComfyUI-Florence2
FAQ
Comments (10)
Error occurred when executing LoadImagesFromDir //Inspire: Directory 'D:\VER_13\_AI\ComfyUI\ComfyUI\input\input_supir cannot be found.'
Yes, that directory path lead to an image file used in my example showcase. You need to enter a path leading to a file on your system for it to work.
@RedPinkRetro will it work with images sequence?
@toteche927 If you set "Prompt SWITCH" and "Tag SWITCH" to Florence 2 and WD14 respectively, you would only need to increment the "Selection" node found in [Image Source — Folder] to go through a directory one by one.
I plan on updating this workflow soon and can build in a way to do that conveniently. Thanks for your input! 🙏
Very nice workflow! And cool colors=)
Btw, can you tell me how you colored the node headers separately from their body? I only know how to repaint them entirely.
Hey there, thanks!
The coloring was broken after some update and I could not find any viable alternative, so I contacted the author of this node pack : https://github.com/blib-la/blibla-comfyui-extensions and ended up adding a PR that adds Custom Title, Custom BG and Custom All to the context menu... I am very particular about my colors 😅
Coloring multiple nodes at once only works through a workaround though unfortunately, by moving all nodes that are supposed to share a color into a temporary group and then coloring all nodes' title bars that are within that group. Silly and inconvenient, but it works.
@redpinkretro Thanks for the reply. Yes, indeed, it is strange that many interface solutions, such as color selection or others that we are used to see in other software interfaces, have not yet been implemented in Comfy, which makes it still a tool for those select people whose brains are sharpened for programming and so on. In addition to flexible color settings, there is a great lack of some sane button to save images instead of all these annoying save nodes. And also the file naming setting, which does not need to be written with your hands in a small line where it is impossible to see anything. Oh, I can lament about it for hours, sorry me=)
@13_february Hey there, excuse the very late answer. I was never notified of this message nor of your kind review, thank you very much for that! 🙏
Yeah, you are absolutely right about that. Considering how many node packs implement more or less the same basic nodes alongside their actual projects, just to fill the still existing gaps of core ComfyUI.
I guess everyone is too busy with more pressing matters coming up in a continuous stream as AI progresses so rapidly.
I did not expect much to be honest, but I managed to run it and... I'm speechless. I can't believe how well it works! Thank you :)
Thank you so much for the kind words, glad you like it! 🙏