Instagram: https://www.instagram.com/synth.studio.models/
Buy me a☕ https://ko-fi.com/lonecatone
This represents many hours of work that I give away for free. If you enjoy it, Please make sure to 👍like and post to my workflows. That is how you support me!!! Also, feel free to ⚡tip 😉
Simple, easy to use Identity edit workflow:
V6.0
Added a head swap feature
added remove background for significantly better consistency
Allows you to retain the original image size or choose your own
easily switch between 1 or 2 reference images
Includes an upscaler and 🎨 post prodution area.
extensive notes and simple to follow instructions
也有中文说明
Description
added reference sliders for better control over character and image blend
FAQ
Comments (12)
Thanks, version 2 works great
Is it possible to use other LoRas along with the Identity Edit LoRa? For example, character LoRa... Because I wanted to change the LoRa node in your workflow but it didn't work.
Well, you need identity edit for this workflow to work. What you are looing for is a base workflow. The ID edit locks onto an image specifically.
Try my Krea2 simple one. It's in my simple workflows group
@lonecatone23 No, what I actually mean is being able to use other Loras besides the Identity Edit Lora. Because this workflow completely prevents us from using other Loras. If you could enable us to add other Loras, I would be very grateful.
@Enigma38 what are you talking about? There is a power LoRA loader dead center of the workflow 🤔
@lonecatone23 There's only one Lora loader, and I can only load the Edit Lora into it. I wish I could load more Lora. I want to change the node, but I can't figure out where to attach it because I can't move the other nodes :/ Sorry, I'm a bit of a beginner and I don't want to disrupt your workflow.
@Enigma38 Oh!!! Lol, Sorry. I thought you meant the Character Builder one. 🙄
I'll fix it this weekend. I'm in the middle of a big overhaul. This is next
@Enigma38 Added
@lonecatone23 Thank you buddy you are one of a kind :)
the work you've put into it.
I've been testing the latest version extensively, and overall the identity preservation is excellent. However, I'm consistently seeing two issues:
The final output has a noticeably warm color cast. Skin tones often shift toward orange, beige, or brown, even when the source image has neutral white balance and natural daylight.
The image lacks natural micro-texture. While the overall quality is good, the skin appears slightly soft or waxy, with reduced pores, fine skin details, and fabric texture compared to the original image.
I've already tried:
Increasing the input resolution from 512 to 1024.
Increasing the Grounded Encode resolution to 1024.
Lowering ref_boost.
Using different ESRGAN/Ultimate SD Upscale models (including 4x-ClearRealityV1).
Reducing LUT strength.
Using prompts requesting neutral daylight, accurate white balance, realistic skin pores, and natural photographic texture.
Despite these changes, the warm color grading and soft texture remain.
Are these characteristics inherent to the current Krea2 Edit model or workflow, or are there recommended settings to preserve neutral colors and improve realistic micro-detail?
Would you recommend any specific values for:
ref_boost
Grounded Encode resolution
KSampler settings
Upscaler
Denoise
Any hidden settings that improve texture fidelity and neutral color reproduction?
Thank you again for your work. Any recommendations would be greatly appreciated.
Typically thate Image Edit will grab one or the other as a color grade, so if your image have an overlal warm tone, that will set the theme. I have not noticed that.
Two options:
Option1: Place an Image adjustment node between the vae and final output (preferred). Run a small test sample using the custom prompts and see how you like it
Option 2: use a color correction node connnected to oteh output and using teh original image (or another ) as a source. I do not knwo if thsi is the way to go as it uses VRAM to mak eteh corrections, and puushing 24 full size pics trhough it may overwelm it. I'd have to test
@lonecatone23 Thank you for your reply and for all your hard work. If you find a good solution or improvements in the future, please let us know. I'd really appreciate it. Thanks again!