Mickmumpitz
patreon
[FREE WORKFLOWS] Generate endless CONSISTENT CHARACTERS from one input image!
🕑 Added 2025-01-20 17:42:52 +0000 UTCUPDATE! February 26, 2025
I just updated the MV Adapter Part of the workflow, so make sure to download the newest versions of the workflow!
This should fix the issues that some people had with the MV Adapter Part of the workflow. Make sure to also check the troubleshooting section in the guide!
----
Learn how to create perfectly consistent AI characters with our upgraded free workflow!
The new Consistent Character Creator v01 makes it easier than ever to generate high-resolution character views from every angle, customize emotions with simple sliders, and place your character in different environments and lighting conditions - all through simple prompts.
Works with both Flux and SDXL, supporting AI-generated and real reference images. The workflow automatically creates Character Sheets in multiple formats, perfect for sharing and iteration tracking. This comprehensive tutorial covers installation, advanced tips and tricks, and teaches you how to train your own character AI model for creating unlimited consistent images for AI movies, comics, and social media content.
Download the free workflow now and start generating professional-quality consistent characters for your projects today!
👉Check out the FREE GUIDE: https://docs.google.com/document/d/1aeVADOeJ5Ke85JzVQjCwGUzs_CFF_P8kdo6iWp8ZVSU/edit?usp=sharing
👉Get the ADVANCED workflows here: https://www.patreon.com/posts/advanced-120405703
👇Download the FREE workflows below!
Troubleshooting
ERROR 1 - Failed to install custom nodes
I had the problem that with the new version of ComfyUI some custom nodes are not installed correctly. This was due to an incorrect version of huggingface hub.
To fix this, open your ComfyUI folder (where “run_nvidia_gpu” is located), click on the address bar and type “cmd” > Enter.
Now copy this command:
python_embeded\python.exe -m pip install --upgrade huggingface_hub
and press “Enter”. When it's done, restart ComfyUI.
ERROR 2 - STILL failed custom nodes
Thank you to my Discord User “Foltin” creating this workaround:
“I think I figured out how to install everything properly. You don't need to use the copy trick in the pinned post.
run update_comfyui_and_python_dependencies.bat
in the python_embeded folder, type “CMD” into the address bar to open the command window and run these:
python -m pip install --upgrade streamdiffusion
python -m pip install --upgrade diffusers
python -m pip install --upgrade huggingface_hub
python -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu124 --force-reinstall
python -m pip install xformers==0.0.29.post2
The important part is that you have to update torch again after updating huggingface_hub and diffusers because they mess up torch.”
ERROR 3 - Could not run 'torchvision::nms' with arguments from the 'CUDA' backend
This might happen with newer ComfyUI Versions while using Face Detailer.
To fix this, open your ComfyUI folder (where “run_nvidia_gpu” is located), click on the address bar and type “cmd” > Enter.
.\python_embeded\python.exe -m pip install --force-reinstall torchvision --index-url https://download.pytorch.org/whl/cu121
and press “Enter”. When it's done, restart ComfyUI.
ERROR 4 - My MV Adapter Output is just a blurry mess
Make sure you are using the correct pipeline:
ERROR 5 - Expression editor not working
You probably just need to put in 1.5 as crop_facter where it says NaN. This is a weird bug that happened after a recent update.
![[FREE WORKFLOWS] Generate endless CONSISTENT CHARACTERS from one input image!](https://img5.patrefans.com/storage/4/la/va/563145-019e8569-da7b-79c3-bb81-b3ac93e29150.png)
![[FREE WORKFLOWS] Generate endless CONSISTENT CHARACTERS from one input image!](https://img5.patrefans.com/storage/5/wg/ow/563145-019e8569-da81-72eb-bf8e-cdd1d65e582e.jpg)
Comments
MrSimmonsSr
I wanted to let you know how incredibly grateful I am for this workflow! It just kickstarted my create endeavors into high gear in a way I didn't think possible! I'm just starting a project to build out a small mobile game and had really been struggling to understand how to make a decent looking character in a way that would eventually allow me to get to the point where I can make simple animations of the character. I need them to be doing simple things in the game, and this is a very robust first step in such a task. Thank you, from a new Patreon supporter! My question for you is if you have any tips (or pointers to other workflows in ComfyUI) that would help me take the character sheets I now have coming out of this flow and use freely available stick figure animations (not sure of the proper term for those) for simple activities like "swing an axe" or "walk" and feed those somehow into an eventual sprite animation sheet? Any clues you could give me on where to look for the next steps of this type of workflow would be most appreciated! I know I need to do something with OpenPose (I think!) and also use some type of sprite screen creator capability, but I'm having a bit of a struggle finding the bits I need and understanding how to wire them all up in ComfyUI, so any tips you care to share would be most appreciated! Thanks again for this awesome workflow!
Mick Mahler
Wait until tomorrow, I will release the next version of this workflow then. I completely reworked everything and I hope it fixes your issues!
Erns
I want to give you an update , I managed with help of chat GPT to get SDXL workflow 250226 running , and yeah it is an interesting experiment, but I don't get anything close to the level of quality you are showing and presenting in the video, I have not tried Flux yet so cant speak for that workflow! also facial expressions are the worst, for some reason it only looks good if I dont have any head rotation those first 3 sliders , otherwise it looks as if front of the face slides off like a jelly from the rest of the head to the side ha ha , but that is nothing to do with your workflow I think that's is just how that node works , it performs best if character looks straight and it does not move head too much ... I prefer not using Florence, unless it is just a head , it always write weird things and gets everything wrong , I prefer using original prompt with another sampler re-run using same seed that created character to fix the issues and correct the image Adapter created in retro character style ... The lighting part was amazing I did not even know that was possible before coming across your video ... there are some great Ideas that I think I can improve on ha ha... I learned a lot and it gave me lots of ideas to experiment with ...thanks I really enjoyed playing and experimenting with this workflow ... next will try the flux one, any particular one you recommend, because there are lots of them
Edwin Spencer
Where do I find these DiffusersMVSchedulerLoader DiffusersMVModelMakeup
paul fedor
having same issue
James Park
what is his discord?
Sérgio Hime
Did someone manage SAVEIMAGE error??
Alex
Return type mismatch between linked nodes: pipeline, received_type(MAKED_PIPELINE) mismatch input_type(PIPELINE)
Alex
Would someone be willing to help me with this workflow? been trying like hell and cant seem to get it to work.
CLAUDE BARAKAT
AutoCropFaces is not compatible with my nvidia 5090, any recommendations?
Harcharan Singh
Hi, getting the MV Adapter error. The process just stops after creating a potrait image
Dániel Dudás
I Got this pipeline error: LdmPipelineLoader An error happened while trying to locate the files on the Hub and we cannot find the appropriate snapshot folder for the specified revision on the local disk. Please check your internet connection and try again. someone who know how to fix this ?
stgy
great video! but i found some workflow is really complicated, and i really hope that you can break it down to build them from scratch. For example, how to build multi-view workflow, how to build upscale, ... and how to connect them to build up a complicated workflow, this would be much easier for beginners, thanks!
Stephan Keya Wilmas
I don't think this is accurate. He compresses a ton of information in his videos and goes through it vey methodically. It's our job to watch and rewatch and understand it little by little. There is lots of support in Discord. You have to put in the time to understand this stuff. It's complex and he has made it as easy as possible but we still have to do our part. I also recommend not trying to set this up on your computer as it's very intensive and using something like runcomfy.com.
Stephan Keya Wilmas
Did you get the docker? I just joined the discord but I don't know where to find it!
Edward Berger
I am using MimicPC and it looks like it works very well.
Mick Mahler
Yes I used it there once! Only the "color match" node did not work for some reason so I had to bypass that.
Mick Mahler
Yes I plan on doing that in the future! I am currently building a team to help me with stuff like that. Someone from my community already created a docker for this workflow and it works really well. Are you on my discord?
Alexandre
Hey! Has anyone made this work on Think Diffusion?
Dávid Szalai
Hey Mickmumpitz, First of all, thank you so much for your amazing work on creating consistent characters! Your dedication and tutorials are truly inspiring. I'm struggling with setting up everything needed to run the Flux workflow. I don't have a powerful machine, so I tried using a cloud service to install ComfyUI and run the workflow there — but the installation process is quite overwhelming, especially with all the dependencies. I was thinking it would be much easier for everyone if there was a ready-to-use cloud template where users could just spin up a GPU server and start generating images right away. Platforms like RunPod [runpod.io] offer on-demand GPU servers, and it would be amazing if you could create a pre-configured template that people could simply deploy and use without worrying about installation. I'd love to set this up myself, but I'm not experienced with ComfyUI and struggle to get everything running properly. What do you think about this idea? Would it be possible to create something like that for the community?
Guy Bertental
I'm having issues with: UltimateSDUpscale DiffusersMVSchedulerLoader DiffusersMVModelMakeup
Selina Druifmans
For anyone that is reading this: He DOES answers, but he prefers discord. He literally fixed my problems in a few minutes when i asked him on discord... So hop on discord if you have issues or questions!
Frontline Fury
dowload it first, then drag it into the window-that worked for me, but the isssue now im having is i do not know where to get the check poit or the mdel from.
Mick Mahler
Did you use the official contact options in my Discord #need-help channel as described in the Patreon registration? Then I'm sorry if it got lost and I'll be happy to refund it.
Mick Mahler
Oh also are you on Discord? I see stuff much faster there, use the #need-help channel! :)
Mick Mahler
This is very strange. Do you have this issue only with this workflow? It sounds like a ComfyUI error, how are you using it?
Mick Mahler
Did you try the recommendations in the "troubleshooting" section of the guide? Did id produce any errors beside the one you posted. Stream Diffusion should not be the issue here. I am currently updating the workflows since MV adapter pushed an update that changed a lot of things, so feel free to try again
Magali Grande
yes, i've paid and no answer
truzter
Yes, that's true. I'm out as well. I only have basic knowledge, and I can get his workflows to run, but he's not as beginner-friendly as he claims to be. Unfortunately, he presents himself that way. I would ask for a refund if I could. You can only report him on Patreon since he doesn't deliver what he promises.
Andre Prodjo
I cancelled my membership. The creator doesnt seems to care about its own members
上进 王
When I imported the workflow into ComfyUI, the ComfyUI interface appeared completely blank - literally the entire page was white with no nodes visible at all. However, I have already installed all the missing nodes. When I click the "Fit View" button, it throws an error: "Cannot read private member #boundingRect from an object whose class did not declare it". Could you please provide assistance? Thank you.
Alex Atkinson
I got this error Florence2Run probability tensor contains either `inf`, `nan` or element < 0
Andre Prodjo
hi. im newb here. i load the workflow for sdxl simple but the manager cant install node ComfyUI-Diffusers. Help plz? heres the error: Traceback (most recent call last): File "C:\AI\ComfyAI\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\nodes.py", line 2113, in load_custom_node module_spec.loader.exec_module(module) File "", line 999, in exec_module File "", line 488, in _call_with_frames_removed File "C:\AI\ComfyAI\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-diffusers\__init__.py", line 1, in from .nodes import * File "C:\AI\ComfyAI\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-diffusers\nodes.py", line 8, in from streamdiffusion import StreamDiffusion ModuleNotFoundError: No module named 'streamdiffusion'
Magali Grande
⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️Very angry to have paid and have no support, I do not recommend you! I strongly advise people who are hesitating to take out a subscription. And for those who have taken out a subscription, I suggest you show your dissatisfaction via his youtube and patreon comments. ⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️⚠️
Ido Tzfanya
hey where can i get the UltimateSDUpscale missing node ?
Saint66
Hi, I was able to install everything pretty slick, thats a hell of a workflow ;) Anyway, the workflow stops after the 1st upscaling, nothing behind it will be generated (so I just see her upscaled face): Any clues? Right at the start of the queue, I get this errors: got prompt Failed to validate prompt for output 179: * DiffusersMVSampler 1356: - Return type mismatch between linked nodes: pipeline, received_type(MAKED_PIPELINE) mismatch input_type(PIPELINE) Output will be ignored Failed to validate prompt for output 953: * ExpressionEditor 1336: - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' * ExpressionEditor 1337: - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' Output will be ignored Failed to validate prompt for output 1357: Output will be ignored Failed to validate prompt for output 87: Output will be ignored Failed to validate prompt for output 84: Output will be ignored Failed to validate prompt for output 663: Output will be ignored Failed to validate prompt for output 977: Output will be ignored Failed to validate prompt for output 1349: Output will be ignored Failed to validate prompt for output 1157: Output will be ignored Failed to validate prompt for output 1336: * (prompt): - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' Output will be ignored Failed to validate prompt for output 1158: Output will be ignored Failed to validate prompt for output 976: Output will be ignored Failed to validate prompt for output 974: Output will be ignored Failed to validate prompt for output 975: Output will be ignored Failed to validate prompt for output 756: * ExpressionEditor 1338: - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' Output will be ignored Failed to validate prompt for output 1339: * (prompt): - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' * ExpressionEditor 1339: - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' Output will be ignored Failed to validate prompt for output 887: Output will be ignored Failed to validate prompt for output 754: Output will be ignored Failed to validate prompt for output 926: Output will be ignored Failed to validate prompt for output 1338: * (prompt): - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression' Output will be ignored Failed to validate prompt for output 929: Output will be ignored Failed to validate prompt for output 1337: * (prompt): - Failed to convert an input value to a FLOAT value: crop_factor, OnlyExpression, could not convert string to float: 'OnlyExpression'
Agito Rivers
Yes, update ComfyUi to the newest version. If t is still not working, you need to install both manually through Github
Megha Manohara
I believe the issue is from the MVAdapter package. There is a bug open about it - https://github.com/huanngzh/ComfyUI-MVAdapter/issues/43 Upgrading diffusers to the latest seemed to have resolved the issue. pip install diffusers --upgrade Not sure if you need a specific version, but that seems to have helped (no red nodes at least) Make sure torch and xformers also match your OS?Cuda versions too.
さとし / エンターテイメント
i can't install 'DiffusersMVSampler' and 'LdmPipelineLoader', Is anyone else having the same issue?
benoit corvol
You use a vm ? did edit the container space ? i put 400go each myself
Dhanushga Lionel
Hi, I'm having the same issues as well. Did you find a fix for this?
Roid Mommy
This actually helped me a lot already! I was able to make my very own LoRa and publish it to Civit using your workflow!
yannosse
... and now the strangest thing happen.... after installing nodes and updating cuda and nvida drivers to the most recent : I am not even able to open old workflows that were working yesterday. super strange : got this error message : FluxSamplerParams+ No operator found for `memory_efficient_attention_forward` with inputs: query : shape=(1, 4479, 24, 128) (torch.bfloat16) key : shape=(1, 4479, 24, 128) (torch.bfloat16) value : shape=(1, 4479, 24, 128) (torch.bfloat16) attn_bias : p : 0.0 `fa3F@0.0.0` is not supported because: xFormers wasn't build with CUDA support requires device with capability > (9, 0) but your GPU has capability (8, 6) (too old) operator wasn't built - see `python -m xformers.info` for more info `fa2F@0.0.0` is not supported because: xFormers wasn't build with CUDA support operator wasn't built - see `python -m xformers.info` for more info `cutlassF-pt` is not supported because: xFormers wasn't build with CUDA support ........................................... totally don't understand why sudently the RTX A5000 was working yesterday and now isn't >?!? any ideas ?
yannosse
Helo. I am trying this workflow 250117_MICKMUMPITZ_CCC_v01_FLUX_ADV.json , and to begin > the usual missing nodes ;)) ok here is the list : Missing Node Types When loading the graph, the following node types were not found: DetailTransfer easy_cleanGpuUsed BlurImageFast LoadAndApplyICLightUnet ICLightConditioning AutoDownloadBiRefNetModel ExpressionEditor DiffusersSchedulerLoader DiffusersVaeLoader AutoCropFaces RembgByBiRefNetAdvanced DiffusersMVSampler DiffusersModelMakeup LdmPipelineLoader ok... was able to get down to this 2 last ones thats seems to create an issue with my actual versoin of confi-ui latest update done : DiffusersMVSampler LdmPipelineLoader and trying to install those from the missing node ... here is the only one that appears to be missing : (IMPORT FAILED) ComfyUI-MVAdapter Try fix Uninstall This extension integrates MV-Adapter into ComfyUI, allowing users to generate multi-view consistent images from text prompts or single images directly within the ComfyUI interface. ... tried the " TRY FIX", CHECK UPDATE" ... reboot etc ... unfortenately, nothing works. any suggestion ;) ?. thanks.
Daniel
This is crazy value! Thank you!
ZzZzz 🐑 💤
Thx Mick
Delia Suciu
yayy, thank you so much for all your effort ❤️😎
