Inpaint and Supir Upscale Workflow in ComfyUI

preview_player
Показать описание
Download the workflow:

Join and Support me ###

Support me on Patreon:


/ aifuzz

Let’s be Instagram friends:


/ aifuzz1
Рекомендации по теме
Комментарии
Автор

Congratulations, AIFuzz, on creating an excellent tutorial. It's evident how quickly you are progressing. Keep up the great work!

RubenTainoAI
Автор

Amazing! Absolutely Amazing. Just the channel i was searching for. After about a year i think i know i need to stay thank you so much for this.

gainsum
Автор

Thanks for the informative video & and for providing the workflow. ❤❤

lemdanm
Автор

I subscribe. You are the queen of ComFy...👸

oliviertorres
Автор

Love the video and technique here! I didn't know about the auto masking other than Adetailer. I was getting white burn spots in my outputs and realized that the SUPIR SDXL model needs to have a baked in VAE. That fixed the issue for me so maybe it can help someone else also!

demzonemusic
Автор

Thank you! Love the video and the workflow! Like from China

八爪鱼小饼干
Автор

Very nice video and well explained! Thank you for sharing your experiences and for your time. I am a beginner and have trouble downloading the missing files such as nodes, checkpoints, etc. through ComfyUi manager. Could you possibly share the links for the associated files?

fucamhranh
Автор

im new to this so im probably just bad but its hard for to access the required models and stuff i only got the workflow and then a whole bunch of missing stuff in comfy but ill keep trying but just to let you know for those newbies like myself.

clandelta
Автор

My video card is 1080 8GB. How should I set the parameters of the SUPIR Upscale Node to avoid the error: "torch.OutOfMemoryError: Allocation on device. Got an OOM, unloading all loaded models". Thank you very much.

hainguyenhoang
Автор

Wow, can this be turned into a sequence frame? I mean an animation changes into an image frame by frame. That's so cool.

ScottMia
Автор

Great, the results are more precise than in Photoshop with the generative fill, up to what original image size can you use this workflow?

thepuk
Автор

The developer of the ComfyUI SUPIR node is in the process of splitting it out from a wrapper to individual nodes (that allow you to disable stage 1 for instance). Unfortunately since I switched to the developer branch I have the new split apart node and for some reason it doesn't seem like I can revert back. I'm going to wait anyway, since the results for SUPIR w/ that all together node seemed to be inferior to the GRadio demo (see the issues forum on the github repo for a long discussion on this). I came searching youtube hoping to find out how to hook up this new split out version. I have no idea where the tile wires go etc etc. Hoping you do an update when the development branch changes are pushed into main. I guess a heads up that that update will likely break the workflow you have here.

DavidLightman
Автор

If you're doing YT videos, can you please buy a proper suspended mic so that we can hear your voice instead of constant banging every time you hit the keyboard or the table?

contrarian
Автор

I get this error:
Error occurred when executing ESAM_ModelLoader_Zho:

The provided filename does not exist

SmokedHam
Автор

why are you using AI to change your voice 💀

iobqxgr
welcome to shbcf.ru