20251104 update QwenEditUtils 2.2
add Mask option to modify the specific area
https://www.runninghub.ai/post/1985595549766365186?inviteCode=rh-v1279
20251104 update v5.1
add QwenEditAdaptiveLongestEdge to easier adjust longest edge for different image
RH workflow:
QwenEditUtils 2.0 Any Resolution Reference
Experience link:
https://www.runninghub.ai/post/1985595549766365186?inviteCode=rh-v1279
Custom_node:
https://github.com/lrzjason/Comfyui-QwenEditUtils
20251104 update v5.0
20250927 update v4.0
Use with lora https://civarchive.com/models/1939453 v2 for QwenEditPlus Consistence Editing
20250924 update v3.0
Updated plugin for QwenEditPlus
Removed old nodes. The new node should capable old function when disable vl resize.
20250915 update v2.0
Completely solve pixel shift issue.
Use with Consistance Edit Lora, strength: 1 https://civarchive.com/models/1939453/qwenedit-consistance-edit-lora
Use plugin in https://github.com/lrzjason/Comfyui-QwenEditUtils
RH workflow:
QwenEdit Consistance Edit workflow v2.0
Experience link: https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279
Kontext like workflow.
Avoid to encode ref image to qwenvl.
Use with Consistance Edit Lora, strength: 0.48 https://civarchive.com/models/1939453/qwenedit-consistance-edit-lora
RH workflow:
It's extremely fun and I highly recommend it to you!
Workflow: QwenEdit Consistance Edit workflow
Experience link: https://www.runninghub.ai/post/1965033238311915522/?inviteCode=rh-v1279
Description
FAQ
Comments (23)
Good stuff Jason! The custom node pack installs fine but it's not the Advanced one, which doesn't have the VL Resize inputs. Any ideas why?
The lastest repo has the advance node. I just haven't update the readme.
@xiaozhijason I just updated the node pack directly from Github, and it still doesn't show up. The "regular" one does. See here: https://drive.google.com/file/d/1AvXrImMWbAXaFHFVbAo2mGubmMfyY42v/view?usp=sharing
@GordonShock You needs to restart comfyui and refresh web page. If you correctly installed, it would exists. You could see the nodes.py in github. The node is in the file
@xiaozhijason I guess it was the Comfyui manager that fetched a previous version, because direct cloning through powershell did the trick. Before doing so I checked the py.nodes, and the "advanced" one wasn't there. Now it is ;)
@GordonShock I would update the version which might do the tricks to update the fetch list on manager
kinda new to this type of edit usage, are clothes the only thing that can work with this? I can get that swap working fine so thats cool but nothing else seems to be able to swap.
The lora strength is to control the consistency. If it could do the editing without the lora, try to lower the strength. If it is shifted too much, increase the lora strength
does this workflow work in swarm ui too or just comfyui?
I didn't test on swarm UI.
I don’t know why, but the workflow is not loading properly. I have installed the nodes, and they load fine when selecting them one by one. However, when I try with your file, everything appears gray and nothing works. Maybe you need to upload a newer version.
I don't know. Might be the comfyui version is too old or too new? I dont think it could be fixed If I upload a new version. You might build the workflow by yourself after installed the plugin
I got the same issue (I think). Can not load LayerStyle. Uninstalling / reinstalling did not fix it
@svooooosh try to update all your comfy nodes, i did it and loaded.
I love this workflow, it just works. I still don't know where are the images stored, it seems that they are not saved, looking for a solution, good stuff brother!
Nvm brother, i bypassed the comparisons and the reel, and use save image from the vae decode, it works fast as hell! Thanks!
Very simple overlay clothes on a body. When i describe that the person has to wear this on the reference picture, he will only overlay it over the picture. So the person not wear it. Most times a picture in in the background. Second problem is: It will make the skin more plastic bad. Consistence not really. What you can do is put faces merge together or very similar objects. There are better solutions for clothes and i think face merges too.
It is a simple tool lora. It doesn't change how the model perform. If the original model perform good result, you could just turn the lora strength lower. If the original model completely failed the task, this lora couldn't help. The specific lora for clothes and faces definitely perform better than this lora. The main purpose of this lora is to prevent pixels shift.
I did some changes to your wf so i can just edit one picture with a prompt, without a second input image and maxed the output resolution. Damn, this is so much better then the original Image Edit WF without the Lora :O Very consistent faces, absolutely amazing!
Hi there, would you mind sharing your said new and improved workflow? Thanks a lot
Would be great if you can share the improved workflow
I just did it with one image as well. Faces were spot on. Very nice. To just use one image take out the second Qwen Edit Config Prep. As well as the second load image.
same here if you can upload an image using your version of the workflow that would be appreciated if its not too late to ask