I tried it with Machine Delusions' new Animateddiff LCM model, and got the best results so far! Thank you soooo much!!! The IP adapter images really came through fabulously...
Fantastic Tutorial as always bro! Once again huge thanks for all of the hard work you and the crew pour into crafting the workflows and instructional share outs. The LCM is something I’ve been meaning to give some sessions to and BOOM here you are with another dialed in workflow. ⚡️⚡️⚡️
Okay, I ended up having to do a clean install because my ComfyUI got messed up so bad while I was trying to solve all the errors I was getting. Inner Reflection's model links are out of date and coming up 404. Is there a place that has all of these resources in one place or at least clearly defined so I can get the right ones? I'm not well-versed enough to hunt for them. I tried but there are a lot of different files that are named the same thing and when I've tried to choose, I get errors. Thanks.
Im getting the following error when I queue prompt: AnimateLCM_sd15_t2v_lora.safetensors' contains no temporal keys; it is not a valid motion LoRA anyone knows how to fix this?
Same here... I am not sure it is a valid file... for now using a different V1 model (with errors but works) and likely the ones from AnimateDiff itself (like Pan, Tilt, Zoom etc) work. If you figure out how to fix the AnimateLCM Motion Lora error, let us know :D
as always, great and detailed tutorial bro, one issue , i was unable to load the json file onto comfyui, there was no error or anything, just a blank screen, i've tried with different browsers but still couldn't load, but my other workflows loads just fine edited : solved, i found out why , it was indeed loaded, but because the nodes aren't set in the default center view, it was wayyyyyyy of to the right hand corner lol
Thank you for posting this note! I am a newb and didn't realize this so I was in the middle of rebuilding the flow from scratch block by block when I stumbled on this and scrolled out, and there it was down there to the right :) I was stuck on creating a separate height and width node for the very first block :)
Looks like one the key nodes is not longer supported. Efficiency Nodes for ComfyUI Version 2.0+: "This node is originally created by LucianoCirino, but the original repository is no longer maintained and has been forked by a new maintainer. To use the forked version, you should uninstall the original version and REINSTALL this one." Not sure what to do
Hey, Great video, can't wait to put it into practice ! Just one question. Where do I put the Clip Vision Model for the IPAdapter ? Do I put it into ComfUI>models>IPadapter or do I out it in ComfUI>models>clip ? Thanks!
I'm still learning ComfyUI but when I downloaded this template and go to 17:38 I noticed that my PRE TEXT and APP TEXT boxes are missing and the manager didn't find them for me. Is there something simple I'm missing? Thanks.
@@civitai That's great news. Thank you. I'm learning by the seat of my pants so some of the instructions I follow are Greek to me until I get more familiar. I really appreciate this tutorial, though. I'm creating an animated movie with puppets, miniatures, and 3D models and I really want to have the characters be SD generated. This seems to be the magic bullet for it.
@@civitai But now I have to figure out why the Lora Stacker isn't seeing my models, even though the extra model paths do seem to be connected correctly. I'm the worst at this stuff but I'm determined.
Okay, I solved that problem, too. This is intense. I'm still going through each step to make sure I've got everything right before I've generated anything. I really want that face swaper to work. Cheers.
Hi, I'm having a few dependencies issues, even after using install missing nodes in comfy manager. 1)I cant find the lcm >> sqrt_linear beta scheduler anywhere, is there a place i can download? 2) When i load the json file it says its missing depthanythingpreprocessor. I downloaded the depthanything file and placed it in my controlnet folder? In effect one of my nodes in the Depth Controlnet group of nodes is red. and finally 3) It says im missing ADE_loopedUniformContextOptions, which im guessing the ContextOptions Loop node? That is red for me in my workflow as well. Thanks for the help!
Weird I’m getting memory allocation error with a 4090 and over 10GB vram. Gonna try running less frames, but I wonder is it like WarpFusion where a common troubleshoot was to use less controlnets or prompts? I was only using 2 controlnets.
That's the animateLCM motion model. You have to reselect it in your drop down menu and it should work! If you don't touch it, it's going to try to read my directory
Hmmmm, I use 512x896 but I have 24gb of vram. Try running less frames at your current resolution and see if you still get the same error. You may just have to break your runs up.
I try to add a remove baground node, but cannot make it work. i think it could be a great add to your workflow ! Just one question , how do i desactivate the upscaler ? thanks !
Is this the chapter #4 of the beguinners guide? I dont think you have explained how to install comfyUI yet, and I think that comfy UI is a more advances technique than A1111? Please correct me if Im wrong. Thanks a lot for all the videos on your channel Im learning a lot :D
This is a small intermission, part 4 will be uploaded in the next few days. If it's a part of the series it will be in the play list. We're not going into the install for Comfy because it's not beginner friendly. Easy Diffusion and Automatic will be the topics for that beginner series for now :)
Hi great work thanks for your tutorials, i am having an issue when i drag and drop the json file nothing appears and giving me an error of depth anything pre processors and ADE_LoopedUniformContextOptions node missing error
When you land it in, zoom out. The node tree is elsewhere on the screen, sorry about that XD For those nodes, make sure you install the missing nodes in the Comfy manager or Install AnimateDiff Evolved
This is an error some people get that I don't know how to solve unfortunately as I didn't experience it. Hopefully someone can help you out! If not, I'd suggest just scraping the node or disabling the group
I watched this so many times! What a great help. I'm new to Comfy and the road has been rocky. I had one question. Is there a way to generate video that doesn't loop?
Hello and thank you for your great work. 😃 Is there a possibility of making video with Automatic1111 or is it possible just with Comfy ui? Thank you for your answers.
It’s possible by using Deforum and different iterations of AnimateDiff in Automatic1111. Truth is though, the best quality renders for animations will most likely come out of ComfyUI at this point due to it’s flexibility and customizable nature 🙏🏽🙏🏽
I just went to my profile and downloaded the exact file I uploaded and dragged the json into the ComyUI window to see if it was indeed broken. Everything was connected and everything worked right from the jump. Also no one else has experienced this. I'd suggest installing whatever nodes you're missing or trying again.
We confirmed as well, the workflow as it's downloadable on the website, works once you drag the json into Comfy and install any missing Nodes via the Manager. If ReActor isn't working for you, just delete that node group as per the tutorial.
@@civitai Amazing workflow, and thank you, but on my computer i had to re-connect the bottom (4th) "Prepare Image for Clip Vision node" to its corresponding "Load Image" node.
Thank you for this tutorial. Learned a lot. Having a problem with the motion LoRA from LCM Hugging Face ... ComfyUI says it is NOT a valid Motion LoRA :( """ ValueError: 'AnimateLCM_sd15_t2v_lora.safetensors' contains no temporal keys; it is not a valid motion LoRA! """" Any ideas? It works if I use a V1 motion lora (with errors but it still goes though) Thanks again
So, I am struggling to get this totally working. I make it to about 65% green bar, and it errors out. Seeing problems with the IP Adaptor; here is my error code text: Error occurred when executing IPAdapterApplyEncoded: InsightFace must be provided for FaceID models. File "C:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\execution.py", line 152, in recursive_execute output_data, output_ui = get_output_data(obj, input_data_all) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\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:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\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:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_IPAdapter_plus\IPAdapterPlus.py", line 698, in apply_ipadapter raise Exception('InsightFace must be provided for FaceID models.') Anyone have any ideas or get a similar error? Thanks!
ok, so I am a dummy.. I did not have the right IDAdaptor bin file. Updated that, and now I am making it to 90% green bar and crapping out at section 6 KSampler / Hi res fix section with this error: File "C:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch n\modules\module.py", line 1511, in _wrapped_call_impl return self._call_impl(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch n\modules\module.py", line 1520, in _call_impl return forward_call(*args, **kwargs) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\AI\ComfyUI_windows_portable_nvidia_cu121_or_cpu\ComfyUI_windows_portable\ComfyUI\comfy\cldm\cldm.py", line 295, in forward assert y.shape[0] == x.shape[0] ^^^^^^^ AttributeError: 'NoneType' object has no attribute 'shape'
@@todddyer6946 check torch see if any models or lcm models not insert correctly , go over thro the video , u may find some small part that missed read , if u fixed it , should works fine