Hello. I get the following error :( Error occurred when executing IPAdapterUnifiedLoader: ClipVision model not found. File "C:\Users\xxr\Desktop\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\execution.py", line 152, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\xxr\Desktop\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\execution.py", line 82, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\xxr\Desktop\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\execution.py", line 75, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\xxr\Desktop\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 559, in load_models raise Exception("ClipVision model not found.")
@Beauty.and.FashionPhotographer10 сағат бұрын
how do you trigger the prompts to come out/be generated . you clicked a queue button on the top right, which (naturally) does not exist in comfyui . yet the queue button im manager does exost , and when clicked it triggers the workflow to start , and not the prompts to be generated with florence, .... (yet again: naturally)...
@PixelEasel8 сағат бұрын
it is the new interface. update comfy and the manager and you'll see it
@Beauty.and.FashionPhotographer7 сағат бұрын
@@PixelEasel yeah...no instructions to do this,... so it works..... the old manager says a restart must be done to get it updated in the manager UI.... yet a restart doesn't do the update...(naturally yet again ,...something doesnt work ....).... but the bigger challenge is, that i get black images, after i downloaded so many of these files not yet on my ssystem,... and updated nodes and so on ... so now it finishes the workflow to the end , no warning window anymore.... but .... with black images...... (now be honest,... why am i not surprised ..... )
@nobody-u-know19 сағат бұрын
I see the instantid/diffusion_pytorch_model, but not the SDXL/diffusion_pytorch_model. Where can I get that? Are these the same thing?
@gnsdgabrielКүн бұрын
Very good, thanks.
@giusparsifalКүн бұрын
Hello and thanks, I think I have a problem with diffusion pytorch model, can you tell the weight? Thanks
@PixelEaselКүн бұрын
thank you! can find the link in the description
@shshsh-zy5qqКүн бұрын
I love your tutorial. it is well structured and explained well. thank you! subscribed!
@PixelEaselКүн бұрын
thank a lot! happy to hear that you liked it!
@Pikselation_T.I.W.Күн бұрын
And people say A.I. art is not art. This is astounding, and such a great video by the way.
@PixelEaselКүн бұрын
thank a lot! and great comment by the way : )
@jamphire_Күн бұрын
*AI could be art* everything can be art. AI is a tool like other tools, and half of the discourse is understandable because that's what happens with new tools. When the camera was invented, many painters protested that cameras couldn't capture artistic vision, but we can see how far photography pushed art culture forward and even gave painters an opportunity to leave realism and look for new ways of experiencing paint, like impressionism and others. But what people didn't realize was the other half of the protest, something that to this day goes under the radar, something that even now criticizes AI art. It's the simple fact that just because these tools are getting easier to use doesn't mean they magically make you an artist or the work art. What painters feared back then was that new students would stop understanding art and stop reading technical and historical things about art, and just every lazy person who found a camera and took a picture of their cat would claim to be an artist. _I don't go into detail but even this became a discussion point in those days things like why this is not art why something your grandma paints should be held lower than a work by a famous artis and so on_ In my opinion, part of the art's value is the human passion and hardship behind that work, maybe even the vision or message (or intentional lack of it) makes art art. It's something hard to measure, which is why we humans have different descriptions of art. Every tool that makes it easier to create art makes it harder to feel the hardship and the passion behind that work. I feel that part of the appeal is experiencing the inner mind of fellow humans or animals. So, a small portion of the protest against AI is from those who love art with passion and put it in a high place. When they see someone with empty mind copy-paste a prompt that makes an image of a sexualized, high-contrast, low-effort of a female, they get offended. Usually, passionate and mindful people find others like them fast, and they can feel those who are not like them just because they can see opportunities and routes that others don't, so it makes it hard to tolerate those who are not. And I end with this: your grandma might not know about art composition or the golden ratio, but if she does a job containing those elements, it just shows that without knowing what they are, she can feel the roads of art and likes to wander in them. Perhaps being a wanderer lost in those roads is better than a person who searches KZfaq for how to draw a perfect, pleasing eye, just because one is trying to find inner self and its interaction with patterns of a patternless universe, while the other is just looking for attention.
@Pikselation_T.I.W.Күн бұрын
@@jamphire_ Yes! Indeed. Someone who gets it! ❤
@jamphire_Күн бұрын
@@Pikselation_T.I.W. ❤tnx beautiful
@Kreative3DMedia2 күн бұрын
One of the best SD content creators out there.
@PixelEasel2 күн бұрын
thanks 😊 🙏
@mattm73193 күн бұрын
Good info and quick to the point, another great video!
@PixelEasel3 күн бұрын
thx man!
@HereForASeason3 күн бұрын
How are you running the queue without the other ksampler firing? Are you bypassing it but just not showing that step in the video? I see so many videos where it looks like they're just clicking the queue button and it ignores the entire top half of the workflow.
@PixelEasel2 күн бұрын
If you use fix seed and don't change anything else. it's the same output so the k sampler does not fire
@user-pp9xw9mc5k3 күн бұрын
give me moreee bro more videos ıam waiting
@PixelEasel3 күн бұрын
sub! and you'll get!
@roiter005 күн бұрын
Error occurred when executing LivePortraitProcess: The number of driving images should be larger than the number of source images.
@PixelEasel5 күн бұрын
you should change the number of frames in the frame load node
@roiter004 күн бұрын
@@PixelEasel where?i cant find it
@roiter004 күн бұрын
@@PixelEasel can you uploas screnn of the node with the details?
@roiter003 күн бұрын
?
@roiter00Сағат бұрын
@@PixelEasel WHY YOU DONT HELP?
@hoangcan915 күн бұрын
great video thank you
@PixelEasel5 күн бұрын
great comment 😉
@user-jn5pi6zc6q6 күн бұрын
Hello! I have an error when using your workflow Error occurred when executing LivePortraitProcess: LivePortraitProcess.process() missing 1 required positional argument: 'crop_info' How to fix? Thank you.
@PixelEasel6 күн бұрын
just update the workflow. You can try this one kzfaq.info/get/bejne/mtKkoKuqtN7OgZc.html
@huiAPPOAJ6 күн бұрын
when changing face expressions it is better to use other methods since different expressions change the whole face, not just parts of it
@jensenkung6 күн бұрын
I agree, can you share what the other methods are?
@PixelEasel6 күн бұрын
you can try this workflow kzfaq.info/get/bejne/jN11lqp6p93Xqas.html
@pong4289pan6 күн бұрын
The workflow makes my computer shut down. rtx 3070 ti
@PixelEasel6 күн бұрын
sounds very weird, this is one of the simplest workflow...
@Julienxuxu8 күн бұрын
That's great. But why couldn't I combine the background and vedio in the right place? I mean the vedio does not appear at the position of the mask.
@PixelEasel8 күн бұрын
this workflow is already outdated ... with the updated repo, now u can do it in a different way . I'll upload a new explanation soon
@ramirocoloma80969 күн бұрын
I love your model! have you made it available for us? i would love to make a short film with it! 🙏🙏🙏
@divye.ruhela9 күн бұрын
I get Error occurred when executing ImageResize+: 'bool' object has no attribute 'startswith'
@WiLDeveD10 күн бұрын
Very useful Tutorial. Thanks bro.
@PixelEasel9 күн бұрын
thanks! you're welcome🙏
@michaelbrandonfalk531610 күн бұрын
512x512
@user-pn6ey5dn4y10 күн бұрын
Great video - is it possible to close a specific eye? Or get the model to wink with a specific eye, for example, the left eye?
@PixelEasel10 күн бұрын
you'll need to use the mask to choose the eye you want to close, and if it doesnt work with this workflow you can try this one kzfaq.info/get/bejne/lbKlg7hm3NXYmGQ.html
@user-pn6ey5dn4y10 күн бұрын
@PixelEasel I'll give it a go. Thanks for thr suggestion. I can make one eye shut, but no wink.
@ismgroov409410 күн бұрын
nice sir
@PixelEasel10 күн бұрын
thx!
@michaelbrandonfalk531610 күн бұрын
This is great. Looking forward to trying this out.
@PixelEasel10 күн бұрын
thanks! share your experience
@michaelbrandonfalk531610 күн бұрын
@@PixelEasel trying to resolve LivePortraitProcess.process() missing 1 required positional argument: 'crop_info' I think I need to change my videos maybe
@SedidNewEnvironment10 күн бұрын
I have this error: Lora-Training-in-Comfy/sd-scripts/train_network.py The following values were not passed to `accelerate launch` and had defaults used instead: `--num_processes` was set to a value of `1` `--num_machines` was set to a value of `1` `--mixed_precision` was set to a value of `'no'` `--dynamo_backend` was set to a value of `'no'`
@aimademerich11 күн бұрын
Phenomenal
@PixelEasel11 күн бұрын
thanks!
@ShubzGhuman11 күн бұрын
im the first buddy
@PixelEasel11 күн бұрын
cheers! 🍻
@ShubzGhuman11 күн бұрын
Error occurred when executing MaskBlur+: object of type 'float' has no len() File "C:\Users\shub4\Downloads\ComfyUI_WORKING\ComfyUI\execution.py", line 152, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\shub4\Downloads\ComfyUI_WORKING\ComfyUI\execution.py", line 82, in get_output_data return_values = map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\shub4\Downloads\ComfyUI_WORKING\ComfyUI\execution.py", line 75, in map_node_over_list results.append(getattr(obj, func)(**slice_dict(input_data_all, i))) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\shub4\Downloads\ComfyUI_WORKING\ComfyUI\custom_nodes\ComfyUI_essentials\mask.py", line 44, in execute mask = T.functional.gaussian_blur(mask.unsqueeze(1), amount).squeeze(1) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\shub4\Downloads\ComfyUI_WORKING\python_embeded\Lib\site-packages\torchvision\transforms\v2\functional\_misc.py", line 83, in gaussian_blur return kernel(inpt, kernel_size=kernel_size, sigma=sigma) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\shub4\Downloads\ComfyUI_WORKING\python_embeded\Lib\site-packages\torchvision\transforms\v2\functional\_misc.py", line 110, in gaussian_blur_image elif len(kernel_size) != 2: ^^^^^^^^^^^^^^^^ got this error at the end, using win 11, i5 13th, 6gb vram nvidia,
@PixelEasel11 күн бұрын
try to delete the node. and bring it back again (should fix it)
@ShubzGhuman11 күн бұрын
@@PixelEasel thank u gonna try it,
@ShubzGhuman11 күн бұрын
@@PixelEasel its working thank you brother
@baseerfarooqui589711 күн бұрын
im getting black and white with old artifact images instead of getting coloured images. all nodes are perfectly executed. pls suggest me what is the problem
@PixelEasel11 күн бұрын
u can try with different seed. and u can also try with higher strength with the color controlnet
@hoangcan9111 күн бұрын
Thank you for sharing!
@PixelEasel10 күн бұрын
thanks for commenting!
@BayuRahmawan_id12 күн бұрын
any info for Depthanything.safetensor where to download ?
@SuperCinema4d12 күн бұрын
Good!!!, How to change pose my character? not promt character
@PixelEasel10 күн бұрын
just play with the pose editor
@motherindiafathersurf156812 күн бұрын
INSANE model
@RoshanYadav-v2z13 күн бұрын
❤❤❤
@adriands820713 күн бұрын
Hi, thank you for the video, it is very interesting. I have a problem with the fooocus inpaint models, the inpaint_v26.fooocus.patch is not showing up in the Load Fooocus Inpaint node (I saved it in ComfyUI\models\inpaint as the rest of the models, but this one, .patch is not showing up), do you know why?
@PixelEasel13 күн бұрын
no, but did u try to update?
@adriands820713 күн бұрын
@@PixelEasel I did
@RenderPackRyan14 күн бұрын
What is the situation with all the bypassed (IPAdapter) nodes?
@PixelEasel13 күн бұрын
if you want to give reference to the model
@sarpsomer14 күн бұрын
Smart approach!
@PixelEasel14 күн бұрын
smart comment😎
@asyrafyahaya789014 күн бұрын
why sometime when i click generate, sometime the face detailer like ignore to fix face and just continue to my next nodee which is upcaler (without fixing the face first)
@PixelEasel13 күн бұрын
not sure, maybe it has difficult recognize the face due to low resolution
@knightride963515 күн бұрын
Beautiful, thank you
@PixelEasel14 күн бұрын
thanks!
@sarpogartis15 күн бұрын
outdated content
@PixelEasel13 күн бұрын
you can try this method for inpainting kzfaq.info/get/bejne/lbKlg7hm3NXYmGQ.html
@shuanshuanzai15 күн бұрын
Can you refrence the ControlNet model?
@PixelEasel13 күн бұрын
for 1.5 huggingface.co/lllyasviel/ControlNet/tree/main/models for sdxl huggingface.co/xinsir/controlnet-union-sdxl-1.0/tree/main
@ZedZero-z7e15 күн бұрын
Great write-up, but something broke since you posted it. I keep getting the following on a fresh install of ComfyUI and all the nodes you suggested. Error appears in training, labelling works just fine: ModuleNotFoundError: No module named 'accelerate'
@PixelEasel13 күн бұрын
not sure why, but pay attention to the folder name and the path of the folder
@woolenwoods66516 күн бұрын
amazing tutorial, however , it is a little bit unfriendly to beginners who haven't used the nodes you mentioned in the video. is it possible to deconstruct this complex workflow with short episode ? still thanks for your video.
@PixelEasel13 күн бұрын
thanks! ill do my best to make it clearer next vid
@pressrender_17 күн бұрын
Thanks a lot for this tutorial. I wasn't able to activate IPAdapter, I don't know why. Error occurred when executing IPAdapterAdvanced: Error(s) in loading state_dict for ImageProjModel: size mismatch for proj.weight: copying a param with shape torch.Size([8192, 1024]) from checkpoint, the shape in current model is torch.Size([8192, 1280]).
@pressrender_17 күн бұрын
I think I loaded a mismatched CLIP vision model XD
@PixelEasel16 күн бұрын
is it work now?
@Milo_Estobar17 күн бұрын
Downloading controlnet models are easy.... The hardest part is getting Fannovel16's auxiliary controlnet to work especially the *preprocessor*... First of all the ComfyUI's preprocessor folder are not centralized... Example: 1. depth = ...\custom_nodes\comfyui_controlnet_aux\ckpts\LiheYoung\Depth-Anything\checkpoints 2. depth indoor =...\custom_nodes\comfyui_controlnet_aux\ckpts\LiheYoung\Depth-Anything\checkpoints_metric_depth 3.Openpose =...\custom_nodes\comfyui_controlnet_aux\\ckpts\lllyasviel\Annotators <---- then inside this folder are mix with any Illyasviel's works So, every preprocessor has different folder format depending on the author... Preprocessor folders are a mess... While controlnet models folder are centralized with other models like base model folder... Having easy path means we can manually download and put them into into their respective place if ComfyUI downloader not working... Also ComfyUI are really crap at downloading, bad request and shit... which adds another problem for downloading preprocessor... So far I've been using inpainting and image reference to modify a character image... Its feels like going to work with bicycle while everyone using exotic cars.... T.T
@SuperSSystem20 күн бұрын
This video really takes you from 0 to hero, thanks for the good work. I have installed all the links that you provided, got latest Py and Git and all those missing nodes.. just one thing Error occurred when executing MaskBlur+: object of type 'float' has no len() ControlNet works with lora256, I can see the depth mask. In the Mask Refinement Section, the Preview Bridge and the Image Preview are working. IpAdepter Nodes are faded. SegAny section is also OK. The end (Preview)Images are showing, but the dress in both is marked with gray. any suggestions..
@PixelEasel16 күн бұрын
first of all, thanks! not sure what the reason - maybe try to reload it after you got all nodes installed
@PixelEasel20 күн бұрын
I came across this on the MimicPC site and found it interesting. Click the link to generate your workflow and share it in the comment area. You'll receive a $5 free usage quota and also have a chance to participate in the lucky draw to win LTD rewards by sharing your work! Workflow link: home.mimicpc.com/app-image-share?key=19e54391477c4f7e9d2fbfede5ad44e3
@ismgroov409420 күн бұрын
thx ... but..
@windows13zumba-od7lu21 күн бұрын
tnx bro. Can you make a video about weights in SDXL prompt?
@user-ip4my7yi9l21 күн бұрын
very well explained, awesome workflow. 10/10 video. extra points for sharing the workflow for free so i dont need to pause a billion times to recreate it myself and can just look at the workflow myself to learn it.