Home
Creators
Search Recent Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

ULTIMATE ALL-IN-ONE FLUX WORKFLOW V2! (Patreon)

Published:
2024-11-02 18:30:44
Edited:
2024-11-20 21:42:24
Imported:
2026-06
Tags:

Downloads

Content

!!PLEASE READ THE POST BEFORE DOING USING THE WORKFLOW!!ALSO NOTE: After the missing nodes download, a white screen will appear, just press F5 to refresh the browser window and it will look like it's loading for a while, that's normal, it's just downloading a bunch of stuff and additional models in the background.

It will then give a message like a bunch of nodes are still missing but when you go inside the install missijg nodes, nothing is showing. That's ok, it's just comfyui being dumb.

Just re drag and drop the workflow json file and the message will disapear

Hey everyone! I've created a 1-click installer for the newest my "ULTIMATE ALL-IN-ONE FLUX WORKFLOW V2". NOW WITH PULID, Face swaps, Automatic face and hands fixer, Outpainting and even image to video generation!!!! It's really really cool!Hey everyone! I've just released a new video on my "ULTIMATE ALL-IN-ONE FLUX WORKFLOW V2" that is simplified and optimized for complete beginners and people coming from the automatic11111 webui that have never used comfyui before. NOW WITH PULID, Face swaps, Automatic face and hands fixer, Outpainting and even image to video generation!!!! It's really really cool!

You can check out the video right here: https://youtu.be/zLGthxUe2jc

I also created a 1-click installer for the Webui for my Patreon supporters for Local AND Runpod, that you can find right here: https://www.patreon.com/posts/115233636

Here the link to the manual install guide: https://pastebin.com/UGpRrq7zDown below you will also find the ULTIMATE ALL-IN-ONE FLUX WORKFLOW V2 for free for you to use, next following updates to the workflow will be for paid members only.

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

PGFX

🙏😊

Wrtk Wrtk

Very thanks you 💪

KizzyMe

thank you !

Ken L. Black

Thanks! Will try later on....

Arggs

Having issues getting the PulID to import via the Manager. It's failing everytime I hit "install missing nodes" is there a way around this? I'm assuming I'd have e to do it manually?

akitaonrails

I am running on Linux under docker (ComfyUI-Nvidia-Docker) and I followed your manual guide for v2 (v1 is working perfectly fine). Even after installing everything, when I try to load the graph it still complains about missing nodes such as workflowsampler and seed upscaler, workflowextra nodes upscaler, workflowflorence, workflowflux controlnet, workflowimage gen sampler, workflowother nodes, workflowparam, workflowsampler img. Is there anything obvious I missed?

Aitrepreneur

I explain everything in the post, basically it's normal. Note: "After the missing nodes download, a white screen will appear, just press F5 to refresh the browser window and it will look like it's loading for a while, that's normal, it's just downloading a bunch of stuff and additional models in the background. It will then give a message like a bunch of nodes are still missing but when you go inside the install missijg nodes, nothing is showing. That's ok, it's just comfyui being dumb. Just re drag and drop the workflow json file and the message will disapear"

Aitrepreneur

Probably because the insightface install didn't work correctly, not sure what installer you used though You can try using the INSIGHTFACE_AUTO_INSTALL.bat, put that bat file inside the python embeded folder, open a cmd window and drag and drop that inside then press enter, this will avoid the cmd window from disappearing so if there are any errors we can see them. This should solve the issue

James Hanggi

Hello, my friend. After installing the new flux and waiting for everything to download I still get this msg1454 (IMPORT FAILED) ComfyUI-PuLID-Flux-Enhanced I have tried to reinstall it several times without success.

eduardo

wierd i dont see the json file i the post

eduardo

oh it showed up right after the coment lol

Bryan Crump

faceswap one face error: Cannot execute because a node is missing the class_type property.: Node ID '#705'

Bryan Crump

faceswap two faces error: Cannot execute because a node is missing the class_type property.: Node ID '#1098'

Bryan Crump

starting to think I missed something because get a similar error on pulid: Cannot execute because a node is missing the class_type property.: Node ID '#1160'

Simon Deng

I love your ALL-IN-ONE workflow so much!

gabbatv

thank you!

eduardo

hey. getting this eror when trying to use the upscaler PyTorch SDPA requirements in Transformers are not met. Please install torch>=2.1.1 how to upgrade tourch in the portable version?

Schrammo

Hey! Thanks for creating this. I do have some problems though. The following 3 custom nodes fail to install: - ComfyUI-GGUF -x-flux-comfyui -ComfyUI-PuLID-Flux-Enhanced I used your installer in a new clean folder to set everything up. Consulted to Github of those nodes and tried to manually reinstall them. This shows up in the console: To see the GUI go to: http://127.0.0.1:8188 FETCH DATA from: C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager\extension-node-map.json [DONE] [ERROR] An error occurred while retrieving information for the 'Florence2ModelLoader' node. Traceback (most recent call last): File "C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI\server.py", line 435, in get_object_info out[x] = node_info(x) ^^^^^^^^^^^^ File "C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI\server.py", line 412, in node_info info['input'] = obj_class.INPUT_TYPES() ^^^^^^^^^^^^^^^^^^^^^^^ File "C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Florence2\nodes.py", line 148, in INPUT_TYPES "model": ([item.name for item in Path(folder_paths.models_dir, "LLM").iterdir() if item.is_dir()], {"tooltip": "models are expected to be in Comfyui/models/LLM folder"}), ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Florence2\nodes.py", line 148, in "model": ([item.name for item in Path(folder_paths.models_dir, "LLM").iterdir() if item.is_dir()], {"tooltip": "models are expected to be in Comfyui/models/LLM folder"}), ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "pathlib.py", line 931, in iterdir FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\\AI\\comfyFlux\\ComfyUI_windows_portable\\ComfyUI\\models\\LLM' Error. No naistyles.csv found. Put your naistyles.csv in the custom_nodes/ComfyUI_NAI-mod/CSV directory of ComfyUI. Then press "Refresh". Your current root directory is: C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI Error. No naistyles.csv found. Put your naistyles.csv in the custom_nodes/ComfyUI_NAI-mod/CSV directory of ComfyUI. Then press "Refresh". Your current root directory is: C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI Error. No naistyles.csv found. Put your naistyles.csv in the custom_nodes/ComfyUI_NAI-mod/CSV directory of ComfyUI. Then press "Refresh". Your current root directory is: C:\AI\comfyFlux\ComfyUI_windows_portable\ComfyUI

Johan Berglin

When trying to load the workflow file I get that these are missing: workflowoutpaintsize1 workflowVIDEOUPSCALER workflowfaceswap2 workflowfaceswap3 workflowREACTORFACESWAP workflowpulidtxt workflowpluidlux

Gerard Colin

Hi there. Nice job you've done here - I like how clean your interface looks. Appeals to my sense of order :) I do have one problem though: Can't get the ComfyUI-PuLID-Flux-Enhanced nodes to install properly. As far as I can tell, it is related to not having Insightface properly installed. I used the bat file provided to do the install, but it is giving me an error message. (I think it might be related to my version of Python). Someone else commented on a similar problem I think. Any advice? RuntimeError: Cannot install on Python version 3.13.0; only versions >=3.9,<3.13 are supported. Thanks

Erik S. Wiig

Manual Install: Missing Node Types When loading the graph, the following node types were not found workflowFlorence workflowFlux Controlnet workflowNodes 3 workflowXLabs Sampler workflowAdetailer workflowoutpaintsize1 workflowoutpaintfluxend workflowoutpaintflor workflowVIDEOUPSCALER workflowvideoencodecode workflowfaceswap2 workflowfaceswap3 workflowREACTORFACESWAP workflowpullflo workflowpulidtxt workflowpluidlux Missing Nodes in ComfyUi Manager show none as missing. I am considering becoming a Patreon supporter, but first I wanted to give this a try.

Seanehty

As a general trouble shooting technique when you are having issues loading models or are missing nodes: 1) Load ComfyUI (Presumes you have run the ULTIMATE_FLUX-COMFYUI-MANAGER_AUTO_INSTALL.bat ;-) ) 2) Clear any loaded workflows 3) Go to the Manager button, click it and ensure you have updated ComfyUI 4) Restart comfyUI 5) Load the workflow you want to use like the V2-ULTIMATE_FLUX_ALL-IN-ONE-WORKFLOW.json 6) Go back to the Manager button, click it and ensure you have installed all missing nodes 7) Restart ComfyUI 8) Clear and reload the workflow you want to use. Then see if you have any errors. ;-)

Glenn Phillip Nilsen

Thank you! Clearing the workspace and reloading fixed it for me.

Daniel

clear your workflow and but the workflow file in confyui again. Then it works

dedydeejays

i have similar problem. sadly after trying what suggested the missing node problem doesnt resolved, can anyone post a list importime for customnodes to check what im missing here

Tommario Minor

I'm getting an error. comfy wont restart. and two customer nodes _lazy_init() File "C:\Users\pasto\AppData\Roaming\Python\Python311\site-packages\torch\cuda\__init__.py", line 305, in _lazy_init raise AssertionError("Torch not compiled with CUDA enabled") AssertionError: Torch not compiled with CUDA enabled

Tommario Minor

also saying import failed for two custom nodes 137 (IMPORT FAILED) ReActor Node for ComfyUI 1454 (IMPORT FAILED) ComfyUI-PuLID-Flux-Enhanced The Fast and Simple 'roop-like' Face Swap Extension Node for ComfyUI, based on ReActor (ex Roop-GE) SD-WebUI Face Swap Extension adapted from https://github.com/balazik/ComfyUI-PuLID-Flux. common fusion methods for multi-image input, some further experimental fusion methods, switch between using gray image (official) and rgb.,

Vivek P

Hi, where is the workflow?

Dalmat Merkuzic

Hello everyone, I spent the entire afternoon today painstakingly installing the Ultimate All-in-One...V2 into my existing ComfyUI setup. The installation itself went shockingly smoothly. But, of course, once it was installed, the floodgates opened, and a pile of issues and errors came rolling in. No surprise there, right? So, I rolled up my sleeves and tackled Python version conflicts, CUDA version mismatches, FFmpeg directory issues, and on and on. Now, I think—I hope—I’ve squashed all the errors and warnings in the CMD window. But after a full system restart, I now have a new, thrilling development: I’m sitting in front of a blank, white screen. Not a button, not a single pixel in sight—just pure, blinding white. Holy cow. Is there any way to disable the loaded workflow or reset ComfyUI? I’m honestly stumped here, folks; I could really use some help. I already posted an SOS in the Discord, but it looks like no one’s got a clue there, or maybe everyone’s still recovering from Halloween, haha.

Kameljonte

Maybe this isn't the best comfort, but I always make a clean new install everytime. Better to have a working Comfy where you download all the loras and stuff. And then install the AItrepreneur-stuff in a new clean separate folder and point the paths using the "extra_model_paths.yaml" file in the ComfyUI-folder.

Dalmat Merkuzic

Hi Kameljonte, I totally agree with you about doing a clean reinstall—usually, that's exactly what I do too. But we’re dealing with software here that weighs in at several tons of gigabytes! If I did a fresh install every time I started a new workflow, well, my SSD provider would be absolutely thrilled. Haha. Anyway, I finally worked my way through it. Sure, it cost me a bit of my life’s precious time, but I managed to fix all the issues. Even the mysterious white screen! Turns out, it had nothing to do with ComfyUI or the workflow itself—just my browser. I switched over to Chrome, and tada! It works! No errors, no warnings, no missing nodes. Everything’s running smoothly now. But for now, I need a beer—I've definitely earned it!

LL

I'm having a problem also with V2. I have downloaded the json but when I load into Comfy, I'm getting the v1 modules. ??? Bad link?

SUTY THE REAL

Open Comfy and drag the new Json into it

Joe Engelmann

love the new version, flawless imo, the only thing i am curious about is has anyone gotten this to play nicely with silly tavern? i always seem to revert back to the silly tavern default using SDXL or 3.5 med because other workflows don't seem to pick up the silly tavern placeholder parameters correctly. any ideas?

Salvador Tropea

About the florence issue: You must create a folder named LLM inside the models folder. Once created the error will go away, but you'll need to download the models. The plugin site is: https://github.com/kijai/ComfyUI-Florence2/tree/main here it explains you can use a node named `DownloadAndLoadFlorence2Model` to download the models, or just use `git clone` for the models you want, I didn't check this part of the workflow, but I think it uses the "large" one. Now about the ComfyUI_NAI-mod issue, this is more complex, the problem is that they changed the name of the project from *ComfyUI_NAI-mod* to *ComfyUI_Universal-Styler*, but forgot to change the code!!!! (I can't beleive it), so you have to edit the *ComfyUI-Universal-Styler/naistyler_nodes.py* file and change *ComfyUI-NAI-styler* by *ComfyUI-Universal-Styler*. This will make the CSV stuff available, you can check it during the load of ComfyUI, it will say the naifilters.csv, naistyles.csv and naitypes.csv are there. You will get the above error when you load the workflow, but this is because the names are hardcoded in the node, and you must change the names (as you change models). I didn't check this part of the workflow yet, but changing the name should be enough. As I'm on Linux I also added a couple of changes to the code, one to change the Windows backslashes (I can't believe the crapy Python let them pass) and I also changed the name *ComfyUI_NAI-mod* to *ComfyUI-Universal-Styler*, so even if a node contains Windows old names the code loads them.

Salvador Tropea

Thanks for sharing your workflow. Here some constructive comments: 1. A monolithic all-in-one workflow is not the best because: 1.1 People gets tons of errors just the first time they try to use the workflow, and they come from parts that they perhaps doesn't even need to use 1.2 A 500 kB workflow means that every single image I generate has 500 kB extra data. In the files I generated with the default prompt this was around 30% of the resulting image 1.3 Now that the new ComfyUI interface is available you can have more than one workflow loaded at the same time, no point in having all the features at once 1.4 I don't see much connection between groups, so having them together is no advantage 1.5 People will use f8 or GGUF, not both 2. The workflow notes should also suggest the number of steps recommended for Schnell model. For people with old hardware Schnell is a very interesting option. You get good images in 4 steps, 1/5 of the time, which is critical if the generation time for 20 steps is 23 minutes. 3. I'm experimenting with LoRAs that reduces the number of steps needed. Did you try them? 4. A nice addition could be AYS scheduler (Align Your Steps from NVidia paper). It helps to get good images in less steps (works for any model) 5. I see people using the A1111 webui talk about 2 guidance parameters, but your workflow has just one. Do you know what is the other? 6. No negative prompt? 7. I know you put a lot of work in the workflow, but the amount of dependencies is overwhelming, I suggest investigating ways to reduce it. Separating the workflow in pieces could also help. The Python AI scene is a real mess the extensions you need installed generate tons of dependency conflicts. People doesn't bother about specifying reasonable requirements, I saw a lot asking for particular versions, either for no reason, or just because they are too lazy to go and change a couple of lines of code. So keeping the workflow dependencies low is important. 8. You are using nodes from ComfyUI-Universal-Styler, but it looks like you used an old version, this creates silly error messages, in fact this custom node is broken

Ahmed Shabana

ya it worked tankx v v v v v much

Laberneth

This part in the manual instruction is unclear to me: then clone this repo from huggingface: git lfs install git clone https://huggingface.co/Aitrepreneur/CogVideo Where? Also into .../models or in the base directory of Comfy? Thank you in advance... PS: Awesome workflow btw. V1 is great, can't wait to use V2. Could take a while due to my slow internet :D

FrozenMetroid

The link is above "V2-ULTIMATE_FLUX_ALL-IN-ONE-WORKFLOW.json"

Lord ArabianShark

Thank you very much; this solved some issues I was having.

SUTY THE REAL

ComfyUI_windows_portable\ComfyUI\models the insightface and CogVideo will be in the models folder. Atleast thats how i understood it and it seems to be working

Joel Roque

Just wanted to say thank you for allowing us to have this great workflow for free.

Laberneth

I fixed my PULID/Faceswap by uninstalling onnxruntime and reinstalling it again. I tried to install it with the *.WHL file, which didn't worked unfortunatly, so i used the basic command to get the latest version. That worked for me... If you want to try it too: Inside the python_embedded folder open CMD prompt and type python.exe -m pip uninstall onnxruntime onnxruntime-gpu Confirm twice with YES and wait until it has finished, then python.exe -m pip install onnxruntime onnxruntime-gpu Also don't forget to click in Comfy on Edit/Clear workflow (Beta UI), load the *.json again and RESTART. Be patient: It will load some files in the background (like Florence) and there will be NO progressbar in the console-window. You can check in your taskmanager (performance-tab/network) if something is downloading atm.

Petri Lummemäki

I understand your concerns. On the other hand, I like the fact that you can experiment with all the tools and get to know them without having to load them one by one all the time, at least at the beginning. I guess there's nothing to stop you from deleting the tools you don't need, saving the file with a new name and then you have a smaller file with just the tool you want. And you can do this for every tool. I didn't realize that ComfyUi stores the workflow and the prompt used in this way inside each image. Thanks for this information. Automatic1111 had a feature that could show the prompt used to make the image and this was very useful if you had a good image and couldn't remember what prompt it was made with or what settings. This information was very useful with ComfyUI, thanks for that.

Johan Berglin

I opened a new clean workflow added one of the nodes and it worked, suddenly later it loaded the missing parts when i reloaded the workflow

Sam Shennan

I'm getting a lot of FaceSwap errors I did a full reinstall of comfyUI from scratch to make sure I didn't bring any issues from my previous installs: ReActorFaceBoost (In group node 'workflow>REACTORFACESWAP') workflow>REACTORFACESWAP ReActorFaceSwap (In group node 'workflow>REACTORFACESWAP') ReActorMaskHelper (In group node 'workflow>REACTORFACESWAP') Int-🔬 (In group node 'workflow>outpaintsize1') workflow>outpaintsize1 JWFloatDiv (In group node 'workflow>outpaintsize1') JWFloatMul (In group node 'workflow>outpaintsize1') JWFloatSub (In group node 'workflow>outpaintsize1') JWFloatAdd (In group node 'workflow>outpaintsize1') JWIntegerMul (In group node 'workflow>VIDEOUPSCALER') workflow>VIDEOUPSCALER JWImageMix (In group node 'workflow>VIDEOUPSCALER') workflow>faceswap2 workflow>faceswap3 Text Concatenate (JPS) (In group node 'workflow>pulidtxt') workflow>pulidtxt workflowoutpaintsize1 workflowVIDEOUPSCALER Crop Image TargetSize (JPS) VHS_VideoCombine workflowfaceswap2 workflowfaceswap3 workflowREACTORFACESWAP CR SDXL Aspect Ratio workflowpulidtxt Do I need to run the models installer bat too, I would have though that wouldn't be gathering these files though.

Melv

the message didnt disappear for me and the nodes are still marked red. They dont show up in the missing nodes of the manager tho :/

Salvador Tropea

I made a tool to extract the prompt and workflow from PNGs generated by ComfyUI, is in Python, for Linux console, but should be usable on Windows: https://github.com/set-soft/Comfy_UI_prompt_extractor The --keep option will keep intermediate files, so even if you have a compressor it will keep the uncompressed JSON files.

SilentMonk

A lot of custom nodes are not installed via the automatic install missing custom nodes. How to fix this? When loading the graph, the following node types were not found workflowFlorence workflowFlux Controlnet workflowXLabs Sampler workflowoutpaintsize1 workflowoutpaintflor workflowVIDEOUPSCALER workflowvideoencodecode workflowpullflo

Paul de lavallaz

Same here. Even after installing all custom nodes, whenever I reload the workflow I get missing nodes error again. Even with nodes that work fine when i open the V1 of this workflow

SilentMonk

For me it works now. Press Clear and then drop de JSON again. Then all nodes load correctly.

SilentMonk

now I am getting SamplerCustomAdvanced mat1 and mat2 shapes cannot be multiplied (1x2304 and 2816x1280) all the time. This should point to mismatching models from different type SD1.5/SDXL/FLUX. However I used all the models specified in the manual, so I don't understand why this keeps happening.

rubz

Also had this for me doing the actions of step 4 of the local install in the right order. i guess clear also works :)

Will Breaker
>>154490823

Thanks a ton for this. Initially I got frustrated and walked away for two days before I decided to look here.

DStiles
>>154490823

Thanks a ton guys, clearing the workflow and redropping the JSON file worked the best. I was going crazy for a minute because the nodes weren't appearing as working. I still got a few nodes that aren't installed, but they weren't related to the PuLID, so I'll figure that out later.

Blair's Secret (edited)

Comment edits

2026-09-08 20:00 This isn't really specific to V2, but PuLID as a whole, every time I try to run this on any workflow I get this message: expected scalar type Half but found BFloat16 I think it's related to VRAM, but I'm wondering if people under 16GB of VRAM can run PuLID at all?
2024-11-07 10:57

Einziger

I have a problem with the Upscaler, it recreate a totally different image with the same composition. I would like the upscaler to upscale my image with fidelity. What I have to do? Did I miss something?

DStiles
>>154593862

Yeah brother, it's a pain to deal with as I never got the chance to test PuLID for myself yet. Hopefully, somebody can help us on this soon.

Einziger

Yes! thank you! Something like this could be a nice addition to your next workflow. I know it's simple but it's something very usefull !

Daniel Garniron
>>154490823

Yes same here, he probably grouped nodes, I noticed I can't load a workflow with grouped nodes .... I dont know how to fix tho

Bunny Swallows

I have never been able to get the pulid to work either. no matter what i do, it just tells me something else is broken, or i have to download something that it won't download.

DStiles
>>154593862

Yeah, I just can't get it to work either, no matter what I do and I can't find a direct solution for it online either which is frustrating. It seems like a lot of people can't get PuLID to work and have just given up on it. I just want to see if it's better than using a LoRA, or not. I think it's something we need to download for sure.

Geralt z Rivii

Installed all extensions but still it says I am missing: 1) flowfaceswap3 2) ReactorFaceSwap 3) pluidlux 4) flowfaceswap2 What I should do? Should I install some extensions manually and which one exactly? I think i even tried to do some additional things for 2) but without sucess I guess.

Geralt z Rivii

Yes. i did whole thing but actualy clear worflow and loading again helped me with this and other missinf.

SlyNine

This fixed my issue as well!

Erik S. Wiig

having problems with image to image, i took a normal photo and ran it throguh image to image, despite the prompt having anime defined, it just recreates a realistic image without any anime art.

hendra alexander

when pulid node i got this error : "pickle data was truncated" please advice thank you.

Len

Thank you for your hard work! Is there any way to output image in specific resolution (i.e. 16:9) in Flux GGUF Image to Image?

Yukon Cornelius

So. Just to be clear. I see a ton of models and stuff for like prompt generation for example. But those aren't actually installed yet right? I notice everything is basically trying to download stuff but it appears to hang up. but i guess in the background it really is downloading. Does anyone know of a way to just make it download everything all at once? maybe then I could let it run overnight and just download everything.

Adam

Do we need to uninstall the previous Flux & Comfy installs? Or can we just delete the old ones?

Starius

Hello! I'm following the installation, however after loading the JSON file, I do not have a "manager" button to click. Any ideas?

Realtoonai

when I run faceswap, I have error message: D:\a\_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:636 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasn't able to be loaded. any idea how to fix this?

Starius

I got that issue sorted thanks to help on the Discord, but now when I try tot use the Queue button I get this error: Failed to validate prompt for output 110: * DualCLIPLoader 124: - Value not in list: clip_name2: 't5xxl_fp16.safetensors' not in [] - Value not in list: clip_name1: 'ViT-L-14-TEXT-detail-improved-hiT-GmP-TE-only-HF.safetensors' not in [] * UNETLoader 123: - Value not in list: unet_name: 'flux1-dev-fp8.safetensors' not in [] * VAELoader 399:0: - Value not in list: vae_name: 'ae.safetensors' not in ['taesd', 'taesdxl', 'taesd3'] Output will be ignored invalid prompt: {'type': 'prompt_outputs_failed_validation', 'message': 'Prompt outputs failed validation', 'details': '', 'extra_info': {}}

LL

I'm confused (which isn't a novel time in my life!) Twice now I've tried to download the V2-ULTIMATE_FLUX_ALL-IN-ONE-WORKFLOW.json from your Patreon sight. According to the screen in your video, I should see "Face Swap person 1,2,3" nodes. I have now downloaded both V1 and V2 and they both open in Comfy with the V1 "Fp8 Image," "FP8 Image to Image," nodes . . . they are the same for some reason. Is it me or bad link? Disclosure: I am running ComfyUI in Stability Matrix not one of your one-click standalones . . .

Mike Cee

im getting this error: C:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\python_embeded>python.exe -m pip install onnxruntime==1.19.2 onnxruntime-gpu==1.15.1 insightface-0.7.3-cp312-cp312-win_amd64.whl Processing c:\comfyui_windows_portable_nvidia\comfyui_windows_portable\python_embeded\insightface-0.7.3-cp312-cp312-win_amd64.whl Collecting onnxruntime==1.19.2 Using cached onnxruntime-1.19.2-cp312-cp312-win_amd64.whl.metadata (4.7 kB) ERROR: Could not find a version that satisfies the requirement onnxruntime-gpu==1.15.1 (from versions: 1.17.0, 1.17.1, 1.18.0, 1.18.1, 1.19.0, 1.19.2, 1.20.0) ERROR: No matching distribution found for onnxruntime-gpu==1.15.1

GlobalGrooveWORLD

WTF??? workflowFlorence workflowsampler and seed upscaler workflowExtra nodes upscaler workflowpositive prompt guidance workflowNegative Prompt workflowFlux Controlnet workflowXLabs Sampler workflowImage gen sampler workflowParam workflowOther Nodes workflowsampler img workflowOther Nodes 2 workflowSampler 3 workflowNodes 3 workflowREACTORFACESWAP workflowAdetailer workflowoutpaintflor workflowoutpaintsize1 workflowfff workflowoutpaintfluxend workflowfaceswap2 workflowfaceswap3 workflowpullflo workflowpulidtxt workflowpullidpara workflowpluidlux workflowvidiupsl workflowvid encod No selected item Nodes that have failed to load will show as red on the graph.

Aitrepreneur

Read the patreon post, everything is explained, that's normal

Daniel Moretto

I have completed the installation, but FP8 Image Generation only generates black screens, without any error notifications.

Bill G.

okay.. spent a real long time loading and finally getting this thing to work yesterday.. used it, it worked, closed it up for night, tried to start program today, and low and behold.. the program start exe file run_nvidia_gpu.bat starts to run and a shit load of errors, stops and (Press key too continue...) then closes.. what happened, as it worked fine after install last night. Figure this one out? It seemed perfect for about 30 mins.. C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable>.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build [START] Security scan WARNING: Skipping C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\colorama-0.4.6.dist-info due to invalid metadata entry 'name' WARNING: Skipping C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\packaging-24.1.dist-info due to invalid metadata entry 'name' WARNING: Skipping C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\requests-2.32.3.dist-info due to invalid metadata entry 'name' WARNING: Skipping C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\tokenizers-0.19.1.dist-info due to invalid metadata entry 'name' WARNING: Skipping C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers-4.42.4.dist-info due to invalid metadata entry 'name' WARNING: Skipping C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\urllib3-2.2.2.dist-info due to invalid metadata entry 'name' [DONE] Security scan ## ComfyUI-Manager: installing dependencies done. ** ComfyUI startup time: 2024-11-14 18:47:31.291588 ** Platform: Windows ** Python version: 3.11.8 (tags/v3.11.8:db85d51, Feb 6 2024, 22:03:32) [MSC v.1937 64 bit (AMD64)] ** Python executable: C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\python.exe ** ComfyUI Path: C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI ** Log path: C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\comfyui.log Prestartup times for custom nodes: 0.0 seconds: C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy 0.0 seconds: C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Easy-Use 2.5 seconds: C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager Total VRAM 16379 MB, total RAM 98206 MB pytorch version: 2.3.1+cu121 Set vram state to: NORMAL_VRAM Device: cuda:0 NVIDIA GeForce RTX 4060 Ti : cudaMallocAsync Using pytorch cross attention Traceback (most recent call last): File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\main.py", line 91, in import execution File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\execution.py", line 13, in import nodes File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\nodes.py", line 21, in import comfy.diffusers_load File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\comfy\diffusers_load.py", line 3, in import comfy.sd File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\comfy\sd.py", line 18, in from . import model_detection File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\comfy\model_detection.py", line 1, in import comfy.supported_models File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\comfy\supported_models.py", line 5, in from . import sd1_clip File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\ComfyUI\comfy\sd1_clip.py", line 3, in from transformers import CLIPTokenizer File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\__init__.py", line 26, in from . import dependency_versions_check File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\dependency_versions_check.py", line 57, in require_version_core(deps[pkg]) File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\utils\versions.py", line 117, in require_version_core return require_version(requirement, hint) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\utils\versions.py", line 111, in require_version _compare_versions(op, got_ver, want_ver, requirement, pkg, hint) File "C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable\python_embeded\Lib\site-packages\transformers\utils\versions.py", line 39, in _compare_versions raise ValueError( ValueError: Unable to compare versions for packaging>=20.0: need=20.0 found=None. This is unusual. Consider reinstalling packaging. C:\Users\BG\Desktop\UI2FLUX\ComfyUI_windows_portable>pause Press any key to continue . . .

tc

I dragged your original V2-ULTIMATE_FLUX_ALL-IN-ONE-WORKFLOW.json into ComfyUI v0.2.7-7-g75a818c, and tried IMAGE TO VIDEO tonight but got these errors: got prompt Failed to validate prompt for output 1035: * ImageUpscaleWithModel 1062:1: - Required input is missing: image * DF_Get_image_size 1062:2: - Required input is missing: image * ImageScale 1062:6: - Required input is missing: image Output will be ignored 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 Output will be ignored Prompt executed in 0.05 seconds I rewatched your video here ... https://www.youtube.com/watch?v=zLGthxUe2jc ... and noticed that at ca. 15:11, your "links" in the video don’t match what I see locally. The workflow seems to fail at the Video Combine node. Here's a screenshot of your workflow as presented in the video: https://postimg.cc/189hw7m3 And here’s what I see locally: https://postimg.cc/RWXrR1hQ

Aitrepreneur

redownload the workflow file, it's a new one, I modified it a few days ago because of an update. make sure that you have "updated all" nodes from the manager

tc

Thanks, K — Your updated workflow fixed it. I know you've been dealing with unnecessary, added pressure because of YT policy idiocy while also working hard to keep us all happy with new info and tech support. Just know: your expertise, kindness, and hard work is VERY MUCH appreciated.

Zoro

when i load up the workflow 2 and drag it in i restart after i download missing nodes then i go to missing nodes and these 2 still appear 136 (IMPORT FAILED) ReActor Node for ComfyUI 1471 (IMPORT FAILED) ComfyUI-PuLID-Flux-Enhanced idk what to do? ive tried to fix it

Brent Bowers

I'm a little confused because I started with The Omnigen 1 click installer which worked as expected. But for the work flow installer When I download V2-ULTIMATE_FLUX_ALL-IN-ONE-WORKFLOW.json I am unsure of where to put it and how to execute it. Double clicking doesn't seem to work. Can you clue me in. Sorry if I've missed this being covered somewhere else.

Bil Mor

All of the workflows work great for me with Q8 GGUF Flux Dev on my 3090ti with 24gb of RAM except for Inpainting: it always hits torch.OutOfMemoryError: Allocation on device Got an OOM, unloading all loaded models. Any advice? I believe you're using the same hardware, so I wonder if I'm missing something obvious? (Same problem happens with GGUF Q4. But it works fine with FP8)

Bil Mor

## Error Details - **Node ID:** 549:5 - **Node Type:** SamplerCustomAdvanced - **Exception Type:** torch.OutOfMemoryError - **Exception Message:** Allocation on device ## Stack Trace

Aitrepreneur

ah funny you say that I was actually waiting for your node to be finished :D Have you improved the vram usage for omnigen? every node that I tried seemed to use more vram than the standalone version with worse results

Salvador Tropea

I meassured less than 6 GB for text prompt only cases and around 6.6 GB for one 1024x1024 input image. My nodes are also using the ComfyUI memory management, so for 8 and 12 GB boards the model will be sent to the CPU space as soon as the image finishes generation (not for 16 or more GB). It also uses ComfyUI VAE machinery (not the hardcoded one), so even in the case that the VAE decode gets out of memory ComfyUI will retry using tiled VAE. Additionally it accepts the FP8 version of the model, so you can save 11 GB of disk space. I also enhanced the model load, is not fantastic, but in my system takes 30% less time. I moved the code to "main" after testing all the original examples, I uploaded results (using less steps to save time) here: https://github.com/set-soft/ComfyUI_OmniGen_Demos/ It works perfectly on my RTX3060, but should work with much less. I added calls to ComfyUI to detect if the inference can be done using BF16 and which device was specified by the user. If something fails should be easy to fix. I also added messages for people trying to enable the "KV cache" when using anything other than CUDA.

Balaji

After updating all the missing custom notes, it still shows some nodes are missing but couldnt see this in manager. For example, getting this in FP8 Upscaler: Cannot execute because a node is missing the class_type property.: Node ID '#308'

Alexandre Portugal

I'm having similar issue, even after re-download/re-install the JSON file @ Comfy-UI.

Aitrepreneur

Latest update a few hours ago broke the image to video node, I'm fixing it right now

TrC450

FAILED: ReActor Node for ComfyUI, ComfyUI-PuLID-Flux-Enhanced These two fail anyone else confirm?

Balaji

Could u let me know what am I missing here, it shows either Null or undefined for the pulid_file. "Prompt outputs failed validation PulidFluxModelLoader: - Value not in list: pulid_file: 'None' not in []"

Aitrepreneur

redownload the newest workflow then re drag and drop it in comfyui

Aitrepreneur

I will update that workflow for the V 2.1 tomorrow

Dan Jones

Nothing's working:

Dan Jones

Figured it out. I think the issue was I didn't see the step in the video, just heard it, so couldn't figure it out for a little while.

Brian Wheatley

After running the installer the UI comes up and I get the message that FluxGuidance and ModelSamplingFlux nodes were not found. I've gone into the manager to look for missing nodes but it doesn't show any there. The command line shows the message: Cannot import ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-PuLID-Flux module for custom nodes: No module named 'comfy.ldm.flux'

Peter Berglund

I wanted to do the inpainting. Loaded an image, made a mask and saved it, but it inpaints absolutely nothing – the image stays as is. I tried denoising strength 0.5 and 0.9.

Aitrepreneur

go inside the ComfyUI_windows_portable\ComfyUI\custom_nodes folder and delete the ComfyUI-PuLID-Flux folder, then still inside the custom_nodes folder, click on the folder path, type cmd press enter, this will bring the command prompt window and there type: git clone https://github.com/sipie800/ComfyUI-PuLID-Flux-Enhanced press enter wait for the repo to get cloned, then relaunch comfyui, re drag and drop the workflow and check that all nodes are installed, should work now

Clinton

Have you ever put out a XY plotter for Flux? I have one for SD, but Flux does not work exactly the same way.

Raziel Malichor

Sorry to be a bother. I just installed Workflow V2 and got a message saying "Missing Node Types" with the following things listed: workflowsampler and seed upscaler workflowExtra nodes upscaler workflowFlorence workflowFlux Controlnet workflowImage gen sampler workflowOther Nodes workflowParam workflowsampler img workflowOther Nodes 2 workflowNegative Prompt workflowSampler 3 workflowNodes 3 workflowXLabs Sampler workflowpositive prompt guidance workflowAdetailer workflowfff workflowoutpaintsize1 workflowoutpaintfluxend workflowoutpaintflor workflowfaceswap2 workflowfaceswap3 workflowREACTORFACESWAP workflowpullidpara workflowpullflo workflowpulidtxt workflowpluidlux workflowvideoencod workflowvidsupscal The error message I got in the command prompt is as follows: Error. No naistyles.csv found. Put your naistyles.csv in the custom_nodes/ComfyUI_NAI-mod/CSV directory of ComfyUI. Then press "Refresh". Your current root directory is: C:\1 Flux\ComfyUI_windows_portable\ComfyUI I looked in my folders and the folder "ComfyUI_NAI-mod" is missing from the "custom_nodes" folder. Please help, and just to give a warning, I'm a complete novice at this sort of stuff. So I apologize if I struggle to understand. Thanks in advance

E

everything works but adetailer doesnt seem to work

Raziel Malichor

I watched the video... and I feel kind of silly with how simple that was XD Thank you for your help and the quick reply. Keep up the amazing work, out of all the people that I've seen who do AI content, your videos are without question the best. Thanks again.

E

this is the error im seeing UltralyticsDetectorProvider (In group node 'workflow>Adetailer') workflow>Adetailer workflowAdetailer

Michael Stone

I updated Comfy, and the Queue button stopped working..... it just wouldn't start any jobs. I've attempted a full reinstall. I got as far as importing the .json file, and that's fine. Now I'm attempting to update all the missing nodes, and it's not doing anything. I get the spinning circle in the screen, and nothing happens. The last lines I get in the command window are: FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\832903789_extras.json [DONE] FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\1742899825_extension-node-map.json [DONE] FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\2233941102_nodes_page_1_limit_1000.json [DONE] nightly_channel: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/cache FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\1514988643_custom-node-list.json [DONE] FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\746607195_github-stats.json [DONE] FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\832903789_extras.json [DONE] FETCH DATA from: M:\ComfyUI_windows_portable_nvidia\ComfyUI_windows_portable\ComfyUI\user\default\ComfyUI-Manager\cache\1742899825_extension-node-map.json [DONE] ....and it just stops and does not move. I can get the whole list of nodes by going to "Custom Nodes Manager". That comes up fine. But if I click on "Check Missing", it freezes.

Michael Stone

Update.... I went back to my old install that had stopped working (I had just renamed the directory), updated everything again, and it's working again now. The new install still won't install missing nodes, but whatever was the problem with the old one got fixed in an update.

Flo

Hi, great workflow, I get Failed to validate prompt for output 1218: * CogVideoSampler 1237:1: - Return type mismatch between linked nodes: feta_args, received_type(VAE) mismatch input_type(FETAARGS) * CogVideoDecode 1237:2: - Required input is missing: vae what did i wrong?

Fantasy AI land

any idea why i cant install certain modes, no matter what i do i get missing nodes? missing node types WorkflowFlrrence WorkflowFlux Controlnet WorkflowNodes 3 WorkflowXLabs Sampler

Simeon Hristev

Prompt outputs failed validation UnetLoaderGGUF: - Value not in list: unet_name: 'flux1-dev-Q8_0.gguf' not in [] DualCLIPLoaderGGUF: - Value not in list: clip_name1: 'ViT-L-14-TEXT-detail-improved-hiT-GmP-TE-only-HF.safetensors' not in [] - Value not in list: clip_name2: 't5-v1_1-xxl-encoder-Q8_0.gguf' not in [] VAELoader: - Value not in list: vae_name: 'ae.safetensors' not in [] Prompt outputs failed validation DualCLIPLoader: - Value not in list: clip_name1: 'ViT-L-14-TEXT-detail-improved-hiT-GmP-TE-only-HF.safetensors' not in [] - Value not in list: clip_name2: 't5xxl_fp16.safetensors' not in [] VAELoader: - Value not in list: vae_name: 'ae.safetensors' not in [] Have this 2 error when try txtTOimg and cant find this file. What i can do ?

Travis Little

The "workflowloader" group node will not resolve for me. What are the components that I should have installed that it is built on?

Infinite Jedi

Flux installer no longer working properly

Pertti Ai

I can't see the .json download link :(