Home
Artists
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register
Partychan Matrix ThePornDude
Home Artists Posts Import Register
Patreon importer is back online! Tell your friends ✅

Exclusive Patreon Video: Troubleshooting Flux Workflow Comfyui! (Patreon)

Published:
2024-11-14 18:37:35
Edited:
2024-11-15 01:18:28
Imported:
2024-11
Tags:

Videos

  • 378780438.mp4

Downloads

Content

Video timestamp:

00:00: Intro
01:22: Best practices when creating a folder for Ai tools
02:17: Local install process from scratch
04:45: Initial install complete + verification correct models download
07:45: THE FOLLOWING NODE TYPES WERE NOT FOUND WHEN LOADING THE WORKFLOW
08:40: Manager button not present in comfyui + updating comfyui

git clone https://github.com/ltdrdata/ComfyUI-Manager.git

11:10: Installing Missing custom nodes + error missing nodes
12:50: Error. No naistyles.csv found
13:09: Still missing nodes??
14:00: Magic???
14:20: Change Ui to something more beautiful
15:10: generate Image
15:34: PULID ERROR FIX: TypeError: forward_orig() takes from 7 to 9 positional arguments but 10 were given

git reset --hard cc9cf6d1bd957d764ad418258b61d7e08187573b

17:12: GGUF ERROR Q8 MODELS NOT FOUND
18:08: PULID or Reactor Node missing
18:33: CUDA_PATH is set but CUDA wasn't able to be loaded
19:00: Local install ending info
19:22: Runpod installation
22:49: Loading workflow + Installation of missing nodes
26:06: Conclusion



FIX THE RUNPOD PULID ERROR FIX: TypeError: forward_orig() takes from 7 to 9 positional arguments but 10 were given:

First once everything is installed, go back to the pod and click "Connect to HTTP Service Port 8000" and click on the red stop button. Which will stop comfyui. Then go inside the comfyui folder (where you put the 1 click installer), click on the terminal icon on the right and there type:

git reset --hard cc9cf6d1bd957d764ad418258b61d7e08187573b

press enter and you should see a message saying "HEAD is now at cc9cf6d Rename some nodes in Display Name Mappings (nodes.py) (#5439)"

then go back to the Service Port 8000 and now click on the green button to start comfyui back again, look at the logs files to see when the webui has started back up, it's ready when you see :

"Starting server

To see the GUI go to: http://0.0.0.0:3001"

go back to port 3000 to use comfyui.

Hey everyone! I've created a special exclusive Patreon video on how to completely troubleshoot most (if not all) most common errors for the Flux workflow version and comfyui installation. This video will go through (slowly) the entire installation process from scratch STEP BY STEP for both local and runpod installation. Giving you tips, tricks, advice and fixes to the most common issues, errors and questions I have received since the video came out. This should really help you troubleshoot most if not all of your problems for this particular installation. So, enjoy!

And if you still have issues, don't hesitate to send me a dm with as much info as possible so I can help find a solution to your problem asap ;)

Patreon post with the 1 click installers: https://www.patreon.com/posts/1-click-install-115233636

As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support - now go have some fun😉!

Files

Comments

Eric Dalnes

Thank you! ;-)

John

Awesome, thanks!

LW

Ah... reload the json file... duh! (embarrased laughter... then on to the fun!) THANKS, K

VXLM

Great Video - thanks so musch. Just one more question. I'm using runpod. How do I "rollback" the comfy UI version, so I can use PULID?

Janne Kallio

I am just about to publish an intro video to my Youtube channel about nature, Technology and AI. It is full of stuff done with AI tools I learned from the videos on this channel! Thank you so much! btw I also notice the image to video here is better that Hailuo AI if you want to create video with a person who does not talk. it is then better for LivePortrait. The resolutions for the image to video here are not good though. but I notice today the following: The developers of CogVideoX1.5-5B-SAT have announced that they are currently adapting the model to the diffusers library and expect to complete this process soon. Once the adaptation is finalized, they plan to release the updated version immediately. It will support more resolutions! :)

Maemo

hi , I want to make sure that with the Follower Membership,will i have access to all these videos and one-click installation each month of the subscription??

Glen

No youtube anymore! Lets gooooo 🤪

Bill G.

Thank you so much for this help video was what most of us needed, thank you, it helped a lot, seems to be working now... BUT : I do have one question: When I run the Start nvidia.. "bat" program starts up WITHOUT your V.2 workflow and is standard comffy UI, and I have to load the V.2 workflow manually every time. Is there a way to have the V.2 workload auto load as a default? not sure how to do that?

Aitrepreneur
>>155354783

you can click on the folder icon on the left and load the workflow from there so that it stays linked right there so you can select it whenever you want and switch between different workflows on the fly: https://imgur.com/a/yJTGuvd

Aitrepreneur
>>155328497

First once everything is installed, go back to the pod and click "Connect to HTTP Service Port 8000" and click on the red stop button. Which will stop comfyui. Then go inside the comfyui folder (where you put the 1 click installer), click on the terminal icon on the right and there type: git reset --hard cc9cf6d1bd957d764ad418258b61d7e08187573b press enter and you should see a message saying "HEAD is now at cc9cf6d Rename some nodes in Display Name Mappings (nodes.py) (#5439)" then go back to the Service Port 8000 and now click on the green button to start comfyui back again, look at the logs files to see when the webui has started back up, it's ready when you see : "Starting server To see the GUI go to: http://0.0.0.0:3001" go back to port 3000 to use comfyui.

The Classic

Thank you for this video it finally works when I deleted the folder and started the download process with the video and it works now. Thanks again for this video.

VXLM
>>155328497

Worked. Thanks for your help bro. Awesome.

Matteo Di Lucca

Man, thank you so much for taking the time to put this guide together. I honestly appreciate your effort. No more nodes missing for me! Thanks, my Overlord!

YTJas

Hi! Thanks for providing this workflow. Using Image to Video I get error "Failed to validate prompt for output 1047 * (prompt): - Return type mismatch between linked nodes: images, CLIP != IMAGE * VHS_VideoCombine 1047: - Return type mismatch between linked nodes: images, CLIP != IMAGE" It basically shows me the line connection between videoencodecode to Video Combine has the error. I used installation option 1 since I have 12GB VRAM. Everything else in ComfyUI seems to work.

xxx

Thanx. Special for PULID ERROR FIX: TypeError: forward_orig() takes from 7 to 9 positional. I dont have the problem with my local install which is that latest version (strange) but with RunPOD Install.

Christopher Carlson

thank you, thank you, thank you. not sure what I did differently, but turning off the antivirus I think what made the difference. All seems to work now. Before this video I had given up.

mspi51

Thank you for all of your hard work. I have a new Nvidia 4070 super TI card coming. If I run both a 3060ti and a 4070 super, should I reload the v2 flow and choose option 2 (>12 GB)? Also, is Flux smart enough to use memory from two cards? Thanks again.

Cássio Braga

Congratulations on the hard work bro! You have reignited my desire to use AIs locally

Schrammo

Thank you! I followed step by step and the only thing I changed was the installation path. Previously it was C:\AI\Comfy... and now it is in Users/public/AI... my Username has a space in between. I don't think it should've had any impact on a C:\ installation but it works now so. GOOD JOB!

gabbatv

Hi thanks you for this helpful guide! I am troubleshooting everything but come up with an error i think its the onnx ? can i send you a dm with the video capture?

Elias

Thank you for this video and all your hard work! Very much appreciated.

Aitrepreneur
>>155415002

make sure you got all nodes updated and you have redownloaded the workflow file, I just updated it a few days ago. Definitely try to use the offload to cpu option when using it though, 12Gb is not gonna be enough for full speed generation

Aitrepreneur
>>155419905

Same solution as with local install: First once everything is installed, go back to the pod and click "Connect to HTTP Service Port 8000" and click on the red stop button. Which will stop comfyui. Then go inside the comfyui folder (where you put the 1 click installer), click on the terminal icon on the right and there type: git reset --hard cc9cf6d1bd957d764ad418258b61d7e08187573b press enter and you should see a message saying "HEAD is now at cc9cf6d Rename some nodes in Display Name Mappings (nodes.py) (#5439)" then go back to the Service Port 8000 and now click on the green button to start comfyui back again, look at the logs files to see when the webui has started back up, it's ready when you see : "Starting server To see the GUI go to: http://0.0.0.0:3001" go back to port 3000 to use comfyui.

Aitrepreneur
>>155427841

it's not flux, it's comfyui and it can't use multiple GPUs, so I would recommend going into the device manager and disabling the GPU you don't want to use

Cesar Gaytán

How can I create a character and use the same character for different scenes / images and videos?

brian

I'm still getting the error after trying to install the missing nodes: (IMPORT FAILED) UltimateSDUpscale 136 (IMPORT FAILED) ReActor Node for ComfyUI 1471 (IMPORT FAILED) ComfyUI-PuLID-Flux-Enhanced

pigeon pigeon

my comfyui just says null under the gguf unet model but the model is definitely in the right folder

brian
>>155555848

still not working. i pasted my error text into ChatGPT and this is what it tells me: The error messages you are encountering indicate several issues with your setup for ComfyUI. Here's a summary of the problems and how to address them: Key Issues and Fixes Missing Python Dependencies: Cython and insightface modules are not installed. Fix: bash Copy code python -m pip install cython insightface Run this command in the python_embeded directory of your setup. Path Length Limitation: Windows has a 260-character limit for file paths, causing [WinError 206]. Fix: Enable long path support in Windows by setting the LongPathsEnabled registry key: Open the Registry Editor (regedit). Navigate to HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\FileSystem. Set LongPathsEnabled to 1. Missing CSV File (naistyles.csv): ComfyUI is looking for naistyles.csv in the custom_nodes/ComfyUI_NAI-mod/CSV directory. Fix: Place the naistyles.csv file in the specified directory. Custom Node Import Failures: Some custom nodes, like ComfyUI-PuLID-Flux and ComfyUI_UltimateSDUpscale, failed to import due to missing dependencies or files. Fix: Reinstall the problematic custom nodes by following their GitHub instructions. Ensure required Python packages are installed. OpenCV Guided Filter Issue: The error suggests a missing OpenCV extension (cv2.ximgproc). Fix: bash Copy code python -m pip install opencv-contrib-python Missing Models Directory: The ComfyUI-Florence2 custom node cannot find the models/LLM directory. Fix: Create the missing directory (ComfyUI/models/LLM) and place the required models inside. no idea of any of this is right, or if i should just completely uninstall comfy and try again. i already had comfy and flux installed before trying to use this "oNe ClIcK iNsTaLlEr"

Aitrepreneur
>>155555848

it depends what you have used to install comfy my installer and files are made to work with my "oNe ClIcK iNsTaLlEr" as you say so if you installed it differently before there is no way of knowing what was different with that install. You can try updating comfyui wit the update_comfyui_and_python_dependencies.bat file inside the ComfyUI_windows_portable\update folder. Then launch comfyui, click the manager button and click on “update all”. Once everything is installed correctly, quit comfyui. Download both the INSIGHTFACE_AUTO_INSTALL.bat and the CUDA_PATH_ONNXRUNTIME_ERROR_FIXER.bat files and put them inside the ComfyUI_windows_portable\python_embeded folder, then run INSIGHTFACE_AUTO_INSTALL.bat then CUDA_PATH_ONNXRUNTIME_ERROR_FIXER.bat Once you see the final message : “echo onnxruntime and opencv-python installed successfully” relaunch comfyui and try again. if that still doesn't work then once again this means the initial install might have been different, so you can try using the main installer to install from scratch and follow along with the troubleshooting video.

YTJas
>>155415002

Thank you, it's able to process now. Keep up the good work!

Alex Rodriguez

I finally got everything working. Thank you!

Blue

Thanks for this video. I seem to still be getting a error that is not in the video though. I did a fresh install when I got this error and it's still showing up when I queue the image to prompt section. I checked the LLM folder and the files are there already. DownloadAndLoadFlorence2Model Error no file named pytorch_model.bin, model.safetensors, tf_model.h5, model.ckpt.index or flax_model.msgpack found in directory D:\AI\ComfyUI_windows_portable\ComfyUI\models\LLM\Florence-2-large.

Daniel Kutz

Running first time on RunPod and works great. I do have a question... once I stop the pod, all of my info and the installation is lost? How do I save it permanently for future use? (I tried looking for youtube tutorials but couldn't grasp exactly way it works)

mikrobixmikrobix

Hi, how to get Comfy with Python3.11?