Thank you Olivio for everything you do. Everything you do we can do at home for free. You don't promote websites that we have to pay to use an API. You also don't hide anything behind a patreon only paywall... Thank you so much!
I ran into the same issue with missing nodes even after installation. Manually replacing the modules fixed it for me! Saved me hours of troubleshooting
Didn't he just make an A1111 video before this one? As for comfyui, the spaghetti strings give me anxiety, but I have to admit from everything I've seen it's speed and control after you do manage to configure it is hard to ignore.
A1111 development is currently at a snails pace. There just isn't enough new info to justify more A1111 tuts. All of the new techniques are easier to apply to comfy and that's why it's where the action is.
pip installing without being in an environment could cause a lot of issues down the line. I would always recommend only doing a pip install while you're in your comfyui environment to prevent these potential issues from happening.
I know it is a lot of work making these videos. I would love to see one on how to get the efficiency node to work and get rid of all the red boxes. I have painstakingly followed some of your vids to only get red boxes and even after update I can not get them working. Thanks
Hey this video is not working for many people. Also installing in the base environment breaks previous installs of comfy. Could you please address some of the issues in discord and maybe use virtual environments for future videos like this?
Can't get past missing nodes, Yes I have installed the ComfyUI-3D pack, put the UI claims that Animate Anyone Sampler, Load UNet2D ConditionModel Load UNet3D ConditionModel Pose Guider Encoder Load Pose Guider are NOT found. I have re-started yes. Any advice?
Had the same error and many comments on this video observe this. After research I found a place where it was suggested to simply search for the nodes in the UI/in the workflow and delete the red ones. A reload (double click on the UI and search for the respective nodes) does the trick. I watched Olivios video in detail where he shows/explains the nodes which were red in my UI and just reloaded/inserted them freshly and deleted the old. This worked for me! No need to search for the nodes in the manager or even manually install them!
@@sauronsauronsauronsauron8891 Hi, I tried this, double clicking in the UI and searching for the name of the Nodes, but nothing is coming up. How did you do this?
After installing the missing stuff on comfyui manager, I am getting this error: When loading the graph, the following node types were not found: [ComfyUI-3D] Load Pose Guider [ComfyUI-3D] Load UNet2D ConditionModel [ComfyUI-3D] Load UNet3D ConditionModel [ComfyUI-3D] Pose Guider Encode [ComfyUI-3D] Animate Anyone Sampler Nodes that have failed to load will show as red on the graph.
dude, this random suggestion on the issue solved it for me.... ""I solved my problem, i go to the folder: " F:\ComfyUI_windows_portable\python_embeded\Lib\site-packages" and i DELETE the 2 folders with the name "diffusers" then i go to the manager and i re-install the script, now works!"" basically delete the files, restart comfy, unninstal animateanyone, restart, install from manager, paste downloaded models, restart, works!!!!
@@psaicon0 Tried this. Didn't work. Deleted 'diffusers' and 'diffusers-0.24.0.dist-info' / Restarted / uninstalled AnimateAnyone / Restarted/ Installed Animate Anyone via Manager / Pasted over the files (pre-trained weights and unet) / Restarted / ....same 3-D issue. This comment by @deviltiger78 below fixed it for me: "What worked for me was navigating to the ComfyUI_windows_portable\python_embeded folder then opening a command prompt in that folder and running python.exe -m pip install --upgrade diffusers==0.25.1"
from a comment down below this is what worked for me: What worked for me was navigating to the ComfyUI_windows_portable\python_embeded folder then opening a command prompt in that folder and running python.exe -m pip install --upgrade diffusers==0.25.1
It's totally impressive how you've build the workflow to get the best quality with this addon. I really like the switch you've implemented. I think this is a really interesting new technology and with everything build into ComfyUI it's ready to use, everything customizable of course. Great video about great advancements.
@@IntiArtDesigns You have to rebuild part of the workflow to get it to work. I got it to work but I'm not sure if its worth the trouble, my results are kinda bad, everything melts together
Had the same error and many comments on this video observe this. After research I found a place where it was suggested to simply search for the nodes in the UI/in the workflow and delete the red ones. A reload (double click on the UI and search for the respective nodes) does the trick. I watched Olivios video in detail where he shows/explains the nodes which were red in my UI and just reloaded/inserted them freshly and deleted the old. This worked for me! No need to search for the nodes in the manager or even manually install them!
Had the same error and many comments on this video observe this. After research I found a place where it was suggested to simply search for the nodes in the UI/in the workflow and delete the red ones. A reload (double click on the UI and search for the respective nodes) does the trick. I watched Olivios video in detail where he shows/explains the nodes which were red in my UI and just reloaded/inserted them freshly and deleted the old. This worked for me! No need to search for the nodes in the manager or even manually install them!
When downloading the missing nodes with the manager and then refreshing, it says that the [ComfyUI-3D] node types are missing. Is there a pack that has all of these nodes? This happened when i reinstalled my comfyUI since it was giving me problems last time
Had the same error and many comments on this video observe this. After research I found a place where it was suggested to simply search for the nodes in the UI/in the workflow and delete the red ones. A reload (double click on the UI and search for the repsective nodes) does the trick. I watched the video where the nodes were showing which were red in my UI and just reloaded them an deleted the old. This worked for me! No need to search for the nodes in the manager or even manually install them!
I have tested the template and it works perfectly. The contribution is appreciated. It would be great if you could make a 2.0 version of this workflow adding the possibility of including Lora's and, in addition to the CN DWPose, the HED to optimize the videos.
By the way, the AnimateAnyone-Evolved node breaks your environment if you don't have the exact GPU series as the creator. I would say use chiaojie's nodes instead, or wait for somebody to change the requirements file and remove the extras
Hey thanks for the warning, guess I stay away from this don't want to ruin all the work I put into my environment (again) What are chiaojie's nodes? I get no results on Google :D
@@elowine My bad, it's chaojie/ComfyUI-Moore-AnimateAnyone but these won't work either. You can install them, but without xformers and a correct sdpa implementation in their sampler, it will require 54GB of vram for a 512x768 on AMD.
I've ignored video previously because it still had a long way to go for what I wanted. This is starting to get close so I think I'll take a look Also, thanks for adding the face detailing and video to pose into the workflow. That's really useful and others have been ignoring it.
Also when installing multiple custom nodes, you can use the checkboxes on the left hand side, left of the ID column, once you have selected all the nodes you want, hit install on one of them and all of them will install, saving you time having to do it one by one and waiting for it to finish before starting the next install.
me sale este error, When loading the graph, the following node types were not found: [ComfyUI-3D] Load UNet2D ConditionModel [ComfyUI-3D] Load UNet3D ConditionModel [ComfyUI-3D] Pose Guider Encode [ComfyUI-3D] Load Pose Guider [ComfyUI-3D] Animate Anyone Sampler Nodes that have failed to load will show as red on the graph.
@@IntiArtDesigns If you create the same nodes (their names are in the red part of the failed nodes) and then connect them properly to the other nodes (as they are currently) and then delete the red nodes (original ones), you can get around it.
Hey Olivio, i followed all the steps carefully but a couple boxes wont stop beeing red... these are the onse: [ComfyUI-3D] Animate Anyone Sampler [ComfyUI-3D] Load UNet2D ConditionModel [ComfyUI-3D] Load UNet3D ConditionModel [ComfyUI-3D] Pose Guider Encode [ComfyUI-3D] Load Pose Guider how can i fix this?
Yep, same here, you'd think updating the diffusers models would fix this, but it doesn't. So i'm stumped. And nobody seems to know how to fix this right now. @@magnushhogberg
@OlivioSarikas Hi there. Everything fine and smooth. I have one minor ... not a problem actualy, just - ... the Video Combine dont show the result in the comfy as a preview... it still save the video to the location, but i cant see the result in the comfy. I have TAESD method on in manager, preview enabled in the node, but still no preview shown. When I switch to another of your workflows, the video combine shows the preview there > switch back to this, run again - still no preview... Would anyone know what am I missing here?
i get this error-message: "SyntaxError: Unexpected non-whitespace character after JSON at position 4 (line 1 column 5)" - Can anyone help me how to find the position of the syntax error?
When loading the graph, the following node types were not found: [ComfyUI-3D] Animate Anyone Sampler [ComfyUI-3D] Load UNet2D ConditionModel [ComfyUI-3D] Load UNet3D ConditionModel [ComfyUI-3D] Pose Guider Encode [ComfyUI-3D] Load Pose Guider Nodes that have failed to load will show as red on the graph. Can someone help me?
same.. Cannot import E:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-3D-Pack module for custom nodes: No module named 'diff_gaussian_rasterization' Cannot import E:\ai\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-AnimateAnyone-Evolved module for custom nodes: cannot import name 'PositionNet' from 'diffusers.models.embeddings' (E:\ai\ComfyUI_windows_portable\python_embeded\Lib\site-packages\diffusers\models\embeddings.py) tried installing them manually and it refuses that too because the destinations don't exist
I have a question, its possible to do a 30 seconds video for example? if the densepose duration is 30 seconds? ... I have that problem, i can create only 3seconds videos ..
Is there any way to use lcm lora with this plugin? It is very demanding ad it can be really slow if someone does not have a really powerful gpu. With a 3070ti it will take ~1 minute or more per frame, and even reducing the amout of steps there is no big gain in time
@@OlivioSarikas yeah, moving many items to a separate drive, will copy-paste them back in when needed. went overboard with downloading way too many loras/models early on
i did everything in the video and when i download the missing custom nodes, restarting ui, there r still red boxes, and when i click on queue prompt, i got a KeyError: "class_type" message on cmd, and an unknown error in the ui
Double click one the background, type load image, select the load image node, connect it to the image input that I made between the first and second group, load the image
Thank you for this! I'm just curious if this works for other things like a skateboard trick (with a skateboard) or a boxer throwing a punch combo, etc?
can this be used somehow to just render keyframes or in other words the same character in different poses ? i dont need fluid animation, just wanna render keyframes of selfmade openpose keyframes (exported from blender openpose rig). Or does the model rely on previous frames which are close to the new frame ? ?
This WF requires Comfy UI 3D custom node which is seriusly problematic to install. It does not work via the node manager and manually I failed (yes, user error, but still...)
The file that goes into the clip vision folder when downloaded is a zip file which when unzipped is a folder. Is this normal or does this have something to do with me using a mac?
I tried comfi UI with same prompt minimum nodes no extra close to stable diffusion stable diffusion was lot better in quality n resolution pixel too I just dont know what was the purpose of comfi UI to exist
Wondering if anyone could help me with some info. I use midjourney now, but I would like to do it locally on my own computer. I know we're in the process of this getting better, but I'm mostly interested in consistent characters and video motion added to the images. I'm guessing I use stable diffusion. But do I use ComfyUI? What is A1111? It's all a bit confusing.
First install pytorche with cuda (using comfyui python embeded), next update diffusers to ver 0.26.2 (same way), last when open workflow replace old red nodes to new one (same names). Cant isnert command here, yt delete my answer.
Hey guys i'm new I haven't installed comfyui Yet as I am still learning a111 my question is With this set up Could I add a character I drew and make him Dance?
Some weights of the model checkpoint were not used when initializing UNet2DConditionModel: ['conv_norm_out.weight, conv_norm_out.bias, conv_out.weight, conv_out.bias'] loaded temporal unet's pretrained weights from /pretrained_weights/stable-diffusion-v1-5/unet ... ^C