All things migrate, universal sticker，Support nunchaku
The texture must be a white background image. Just place it where you want
Simple prompt words can lead to better results. It's fine not to write
Trigger word: Put it here
Transparent objects with the prompt word: transparent
LORA:https://civarchive.com/models/1791091
Description
The texture must be a white background image. Just place it where you want
FAQ
Comments (48)
How do i get Image Rembg (Remove Background) to work? I tried pip install -r requirements.txt,tried other versions of WAS Node too but nothing works
Traceback looks like this
Traceback (most recent call last): File "E:\ComfyUI_windows_portable\ComfyUI\nodes.py", line 2124, in load_custom_node module_spec.loader.exec_module(module) File "<frozen importlib._bootstrap_external>", line 999, in exec_module File "<frozen importlib._bootstrap>", line 488, in callwith_frames_removed File "E:\ComfyUI_windows_portable\ComfyUI\custom_nodes\was-node-suite-comfyui\__init__.py", line 1, in <module> from .WAS_Node_Suite import NODE_CLASS_MAPPINGS File "E:\ComfyUI_windows_portable\ComfyUI\custom_nodes\was-node-suite-comfyui\WAS_Node_Suite.py", line 37, in <module> from numba import jit File "E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\numba\__init__.py", line 77, in <module> from numba.misc.special import ( File "E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\numba\misc\special.py", line 3, in <module> from numba.core.typing.typeof import typeof File "E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\numba\core\typing\__init__.py", line 1, in <module> from .context import BaseContext, Context File "E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\numba\core\typing\context.py", line 13, in <module> from .typeof import typeof, Purpose File "E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\numba\core\typing\typeof.py", line 10, in <module> from numba.np import numpy_support ImportError: cannot import name 'numpy_support' from 'numba.np' (E:\ComfyUI_windows_portable\python_embeded\Lib\site-packages\numba\np\__init__.py)
Is there any installed node? Which position reported an error? Just delete the nodes that don't exist. Or replace it with something else.
Futurlunatic I'm using Remove Background from Allab,all good,but i'm struggling with FastCanvas because it's chinese only
Santaonholidays
You can replace it with the "compositor(3)" node of" enricos-nodes", or directly create a graph in PS and paste it in
Futurlunatic Thanks for that,i made a bit minimalistic Worfklow :)
Santaonholidays Congratulations! How effective is it? If it's convenient for you, please help me explain it to everyone. Ha ha
Futurlunatic Its simple,Background stays behind the foreground,you press Continue below the Edited Picture and the result comes out at the Preview Image then you save that and paste it on the main generator
Santaonholidays nice
Nunchaku is a dependency hell, don't install it.
Hahaha, the kontext version is still okay. It's much better than its full version. Who could resist being fast? Hahaha
这流被看着好复杂，对新手不友好
There's no way around it. The texture must have a white background mask. To perfectly reflect it in the workflow, special nodes are needed to assist
I can't use this workflows in ComfyUI, even after install all the missing nodes they still said missing :(
Help me!! Those nodes are missing even though I installed Nunchaku:
NunchakuFluxLoraLoader
NunchakuFluxDiTLoader
Text
TTP_Expand_And_Mask
You can just delete Nunchaku without using it. Delete TTP as well, or find something with the same function to replace it. These are not necessities
Fizee Nunchaku brook :(
I have the same problem, chatgpt helped me solve it. There you need to install the correct version of nuchaku 0.31 and then chatgpt will help you. Send him screenshots and chatgpt will explain to you how to install correctly
Fizee You probably haven't installed the wheels，Among the nodes of the nunchaku, there is one called "nunchaku wheel??" When you run this node, it will be automatically safe and then restart
yuart Thank you very much for your help. You are the best:)
Can you please create a comprehensive tutorial for this and setting it up to work including nunchuku? with links to all required files and how to install?
Tried and tried. No go.
Haha. Please read the instructions carefully. It must be a white background image. A white border will increase the success rate. Simple prompt words will lead to better results
Place your sticker at the position and size you want, completely covering what you need to replace
Never got even close to that step. Endless errors just managing missing nodes.
You can replace the missing nodes or directly upload the pasted pictures and give up using nunchuck
ogbtjyqat450 its the same here.....
endless errors just trying to manage the nodes
Sythexus There is already version V0.2 available for you to try the new version. But your question seems to be a workflow issue
I have an idea. Why not make a lora from that?
Hahaha, there's lora!." What do you mean by lora? My lora is included in the introduction
I wonder if you connect the background image (with no front object) to KSampler latent_image input, and image with front object to ReferenceLatent, will there be any difference in preserving the graphic covered by the white frame? Thanks.
I'm really sorry.." My English proficiency is limited and I didn't understand your question. The best way to use my lora is to take a background image and then use a picture with white borders on all sides as a sticker, placing it at the position you want.
And I have already released version V2 now. You can search for it and use the new one instead of the old one
有使用教程视频吗，大佬，新手看不太懂工作流
B站很多，你可以找下。万能贴，或者kontext最新的迁移工作流基本都是这个。。还有我都已经更新到V2了。。你还在这里干嘛哈哈哈
Why are you loading two loras?? what is actually the second lora file?
Haha, that's an acceleration lora I used during the test. It can be ignored.
flux turbo i guess
I love this... but it doesnt love me.... The node management is like a huge wall which gets taller every time I try to climb over it... I will be on the lookout for the V3
能不能出个工作流说明，还有两个插件怎么都补不齐。
Missing Node Types TTP_Expand_And_Mask
install Comfyui_TTP_Toolset
Restart
Missing Node Types TTP_Expand_And_Mask
install Comfyui_TTP_Toolset
Restart
Missing Node Types TTP_Expand_And_Mask
install Comfyui_TTP_Toolset
Restart
endless
Solution ?
//SOLVED
Install custom node from this link:
https://github.com/TTPlanetPig/Comfyui_Object_Migration.git
JuicyJoy Thanks for your help with the explanation
This lora and the workflow look hugely useful. I have the single image path working great. could someone kindly post the English translation of the FastCanvas buttons? So far I am just guessing. I think that node will change my life once I master it.
Some of these custom nodes breaks my ComfyUI electron build, gets stuck on an endless update loop
Sometimes works, sometimes not. Same input image of your example woman with glasses.
You should remove "anything anywhere" and other exotic nodes (easycleanGpuUsed), it breaks the litegraph arrange layout. I tried v4 and didn"t get better results too.
The workflow is just an aid to me. Specifically, the test effect is still okay. Because I have to test a lot of lora, the global node is used for testing. This can't be helped. If all the wires are wired. I test the workflow will get stuck.
everything is running, but the image generated is randomize not from the source image, i put a white background hand + white bg ring, the results blended together but the ring shape is randomized, not sure whats wrong.