Using the power of Qwen Edit 2511, repose a reference character to a new pose to your heart's content!
All the relevant model DL info can be found within. Minor notes scattered about.
Its very easy to use. Put in a pose image. Put in your character sheet, preferably made with my Lazy Character Sheet workflow. Then just hit run. The loras and 2511 are extremely flexible and it seems a great variety of poses can be executed.
Note that the character in the pose image determines the style, body type and even expression of the final output. If you want a specific style, body type or expression, generate that first.
Version 4.0
Total overhaul
Model is now Qwen Edit 2511 GGUF version
Using the wonderful pose extraction and repose loras from tori29umai (details here)
Simplified everything
Version 3.0
The workflow now uses the Qwen Edit 2509 All-in-One model by Phr00t, found here.
This simplifies the workflow
Also, there's now both a SFW and NSFW model option
Notes updated accordingly
Version 2.0
Now the workflow has been rejigged and no longer uses the repose lora
The model has been updated to the Qwen Edit 2509 model
The workflow now uses the new model's Control Net features using DW Pose
New Image concat outputs an info image with inputs 1, 2 and the output.
Version 1.0
old and can be ignored
Notes:
If you are getting problems like this, try increasing the edge buffer or re-generating the first portion. The edge buffer is by default 5% and the seed is fixed by default.
Cheese and have a good fun genning!
Brie Wensleydale ~
Description
This version does away with the lora by kingroka and upgrades the model to Qwen Edit 2509.
Qwen Edit 2509 comes with Control Net, so I've added a CN DW pose preprocessor to extract the pose for pretty amazing results.
I've also added an image concat feature that outputs an extremely large info image.
FAQ
Comments (15)
It's too bad. it only can generated sfw output..But thanks.
thats why it's too bad?
For now....
Someone will finetune or add a NSFW lora eventually.
I probably don't even need to update the workflow for it to work.
@slipperygem Yeah, I think is more of the model itself. What I wish is not NSFW just more non so plastic look when doing the editing on realistic images but that's also on the model.
@LDWorksDavid This mirror's Wan Video's problem of having anime look 3D the moment there's significant movement or turning.
Its backed into these models, there's probably a lot of synthetic data.
Its the same with relighting, I can't do a relight without making the result look like a pre-rendered cut scene. On that department, it does better with a realistic style.
@slipperygem More than agree. I think they used synthetic data not too realistic as it was "realistic" on the machine.
Братишка оно делает NSFW. You just need use tag "add boy from first person view" add boy to scene.
I will be uploading a new version that switches out the standard Qwen Edit model with the All-In-One V5 model, which has a NSFW version. That version is kind of annoyingly horny, but you'll have the option if that's what you want.
Qwen's AIO model's other benefit is that it has the acceleration, qual, concept loras, as well as the vae and CLIP all baked in, so, its technically simplier, and perhaps slightly faster. Although first model load does take a while.
I'll hopefully put it up today, maybe tomorrow.
@slipperygem >That version is kind of annoyingly horny, but you'll have the option if that's what you want.
This doesn't look like anything to me.
1) Off sd3 latent image its not work with this.
2) use consistence edit lora
Does the consistent edit lora raise the quality? This is a character repose workflow, wo I'm not sure what the point would be. Keep a consistent background I guess?
I think I'll go take a look at the lora regardless though. Thanks.
@slipperygem Consistense keep proportion and hair style on first image. Need doing 0.6-0.8 only.
@meikerAI3d Hey mate, do you have a link to that consistency lora? Appreciate it fren.
@meikerAI3d Thanks mate!