ComfyUI Product Photography Inpainting workflow

Comfy-UI Inpainting workflow for product photograohs
how to take a pack-shot of a real product
and build around it an environment that reacts to it, whether it is with shadow, lighting, reflection or anything else
We will use an IP Adapter V2 if we want to upload an image as a reference
and with ControlNet to maintain the same composition and proportions
well use segment anything to disconnect the product from the original background, and we will composite everything together with the help of Lama remove object node and image composite masked node
#comfyui #stablediffusion #ipadapter #mask #controlnet #segment
follow me @ / pixeleasel
workflow
drive.google.com/file/d/1xXQD...
Segment anything
github.com/storyicon/comfyui_...
lama object removal
github.com/sipherxyz/comfyui-...
juggernaut model
civitai.com/models/133005/jug...
IP-Adapter GitHub
github.com/cubiq/ComfyUI_IPAd...
Prompt Styler SDXL
github.com/twri/sdxl_prompt_s...

[@WiLDeveD]

Thanks bro. very useful content. Please talk more about product involved work flows.

[@PixelEasel]

thanks! i will!

[@aimademerich]

THANK YOU!!! After A year of this amazing tool, finally someone goes beyond and provides some real practical use cases products!!! GOAT

[@PixelEasel]

thanks !! glad you find it useful!

[@knightride9635]

Thank you for sharing the workflow, very interesting

[@PixelEasel]

thank for sharing your thoughts 😇

[@yvrary]

bro this is such a gem account, pls pls pls do more of this but indepth tutorial

[@PixelEasel]

thanks!!!

[@renwar_G]

LEGEND 💎

[@PixelEasel]

thanks !!!😊

[@BoomBillion]

How do you create the different backgrounds using the reference image? What changes do I need to make to the workflow?

[@PixelEasel]

just activate the ipadapter group and you can use a reference image

[@MerlinStark-ig5lt]

Hey, great video! I am relatively new to comfyui. When i load the workflow there comes an error showing me missing nodes in red. Do I have to download them all manually or is there an easier way for that? :)

[@PixelEasel]

thanks! you can use the manager to install missing nodes

[@guillaumewalle]

wow.. i've to test that incredible workflow edit: excuse me, what model should i put in the clip vision for the IP adapter ? i've error when trying using it. otherwise it's crazy !

[@PixelEasel]

thanks! nice to hear!!! you need the model for the 1.5 sd. Also, it's sdxl

[@ai-diary99]

thank you for sharing but on my computer the LaMa Remove Object node not work , do you know how to fix it ? it report err as :Error occurred when executing LaMaInpaint: File "L:\comfyui\ComfyUI_windows_portable\ComfyUI\execution.py", line 151, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\ComfyUI\execution.py", line 81, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\ComfyUI\execution.py", line 74, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-art-venture\modules\inpaint\lama\__init__.py", line 85, in lama_inpaint model = load_model() ^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-art-venture\modules\inpaint\lama\__init__.py", line 53, in load_model lama = load_checkpoint(cfg, files[0], strict=False, map_location="cpu") ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-art-venture\modules\inpaint\lama\saicinpainting\training\trainers\__init__.py", line 29, in load_checkpoint model.load_state_dict(state["state_dict"], strict=strict) ~~~~~^^^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\jit\_script.py", line 874, in __getitem__ return self.forward_magic_method("__getitem__", idx) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "L:\comfyui\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\jit\_script.py", line 867, in forward_magic_method raise NotImplementedError()

[@pvilyou6420]

I have same error, any fix it bro?

[@PixelEasel]

try to update comfy. and check if you download the lama model

[@ai-diary99]

after update comfy ui and this extension,problem still there, i don't know how to download lama model and put where.

[@user-wl1nd3vz8t]

@@ai-diary99 ++++

[@user-ub8zn9sk5w]

Same error. Does anyone know how to download Lama model? Because of my low knowledge, I didn't figure out how to install it from the link in the description.

[@yunongj2487]

Thank you for sharing. This is very impressive🤯! I am new to Comfyui and SD. So far I have got most function goups works, but saw a error in composite group while LaMa remove object run "Error occurred when executing LaMaInpaint: The size of tensor a (600) must match the size of tensor b (2400) at non-singleton dimension 3". Checked customs node and models, everything looks fine. Any sugguestion or quick tensor match feature could be used? Many thanks!

[@PixelEasel]

try to use different resize node

[@yunongj2487]

@@PixelEasel Many thanks! It works perfect!

[@101worlds]

Excuse me, How to create a black mask without white areas inside ?(to make the black more solid). First, I have already removed the background of the main object (such as a cake) and uploaded the image, but the generated mask is not clean (there are white dots inside the black) how to make white dots disappear? thank you!

[@PixelEasel]

you can try playing with the threshold, or using stronger model (in segment anything)

[@chrisgames1831]

Hey bro amazing work. I tried it but it doesnt create the background, it gives me just a grey background. Also (this is probably something wrong with my software, but if you know how to solve it) after loading the workflow and generating 1 image, if I try to generate a second one it says that its running on another tab but it isn't.

[@PixelEasel]

thanks! sorry, but I don't know how can I help

[@chrisgames1831]

@@PixelEasel Alright, thanks anyway 👍

[@lisandroancewicz7821]

it shows this error and I installed all of the models I find with that name: raise Exception("Missing CLIPVision model.")

[@PixelEasel]

make sure that put it in the right folder. I needed to move the old ipadapter models to a new folder for the new version to work I think I explained it in this video kzbin.info/www/bejne/j2bRqamjZct3brcsi=ohrmR-x9Uhpbh-Ck