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

1-Click INSTALL WAN IMAGE AI KING ComfyUI WebUI! (Patreon)

Published:
2025-07-13 19:03:02
Edited:
2025-07-17 18:08:57
Imported:
2026-06
Tags:

Downloads

Content

Hey everyone! I've created a 1-click installer for what might be the CRAZIEST discovery I've made in AI yet, a model originally made for VIDEO that turns out to be THE BEST FREE IMAGE GENERATOR out there! With mind-blowing quality, perfect prompt following, LORA + ControlNet support, and full ultra detailed fast local generation, this model completely DESTROYS Flux and HiDream once properly set up and I’m giving you everything you need to unlock its full potential!

You can check out the video right here: https://youtu.be/oOGiYy7cTFw

The workflow: https://www.patreon.com/posts/wan-image-ai-134046660

The installer of course automates the entire install process for comfyUI and the models downloads!

LOCAL INSTALL!!! RECOMMENDED TO START FROM SCRATCH SINCE YOU NEED THE LATEST COMFYUI TORCH VERSION!!! AND FOR EASY NUNCHAKU INSTALLATION!

Download the WAN-IMAGE-COMFYUI-MANAGER_AUTO_INSTALL.bat

Run the bat file and select either option 1, 2 or 3 depending on your GPU vram and then determine if you want the Original Wan model (recommended) .

Wait for the install to finish

The webui will launch, load the WAN_IMAGE_GENERATION.json file and there you go😎

IF YOU ALREADY HAVE COMFYUI INSTALLED USING A PREVIOUS INSTALLER (not recommended):

Download the WAN-IMAGE-MODELS-NODES_INSTALL.bat and place it in the ComfyUI_windows_portable\ComfyUI folder

Run the bat file and select either option 1, 2 or 3 depending on your GPU vram and then determine if you want the Original Wan model (recommended) .

Wait for the install to finish

???

Profit 😎

IF YOU GET AN ERROR WHEN LAUNCHING COMFYUI AFTER USING THIS INSTALLER SAYING "AssertionError: Torch not compiled with CUDA enabled" it's ok don't panic.

Just go inside the python embedded folder, click on the folder path, type cmd press enter, this will bring the command prompt window and there type:

python.exe -m pip uninstall torch

python.exe -s -m pip install --upgrade torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128

--->replace cu128 by your cuda version (if you don't know your cuda version just open a new cmd window and type nvcc --version it will then tell you at the end Cuda compilation tools, release 12.8, V12.8.93 or something different like 12.6, in that case just use cu126 instead of cu128)

IF YOU ARE USING RUNPOD:

Create an account if you haven't already: Runpod

Click on Pod (on the left side) then click deploy

Choose a GPU with at least 24gb of VRAM (4090 is best), click here for the comfyui template (aitrepreneur/comfyui), then edit the template and choose 80gb for both the container and volume disk, then deploy on demand

Go to my pods, wait for everything to finish and then click "connect", then "Connect to HTTP SERVICE port 8888"

Then go inside the comfyui folder and then drag and drop the WAN-IMAGE_AUTO_INSTALL-RUNPOD.sh file on the left side of the UI then click on the "Terminal" icon on the right side on the UI

Copy and paste these two lines then press enter:

chmod +x WAN-IMAGE_AUTO_INSTALL-RUNPOD.sh

./WAN-IMAGE_AUTO_INSTALL-RUNPOD.sh

Wait for everything to be installed

Go back to the port 3000 and click on the manager to update comfyui and restart it

Load the WAN_IMAGE_GENERATION.json file and there you go😎

Also use the command:

tail -f comfyui.log

inside the workspace/logs to get the real time logs so that you always know when the install is done. It will then ask you to restart, so click on that restart button, wait for the webui to reboot, click refresh once it's done and there you go.

IF YOU ARE USING RUNPOD AND YOU WANT TO STOP THE POD THEN RESTART IT:

Go inside your pod

SERVICE port 8888

Go inside the Comfyui folder, open a terminal icon and re copy and paste those two lines:

chmod +x WAN-IMAGE_AUTO_INSTALL-RUNPOD.sh

./WAN-IMAGE_AUTO_INSTALL-RUNPOD.sh

And then in a few seconds it will reinstall all the requirements for the nodes

Go back to comfyui port 3000, click on manager, click update all, wait for everything to be updated then restart.

Re drag and drop the WAN_IMAGE_GENERATION.json file inside and there you go😎

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

Onyx N

how may GB is the total file size if i use WAN-IMAGE-COMFYUI-MANAGER_AUTO_INSTALL.bat ?

Alex Cooper

Question; I have Comphy already installed from your last files. If I put the installer into a new folder elsewhere on my computer and run it, will that be fine enough or do I need to completely uninstall the old comphy? Thanks in advance

binary

Does it has restrictions like NSFW?

Steve

I you had watched the whole video you would know that you can use NSFW loras with it

Spigott_AI

I just want to thank you for the hard work you’re doing to provide us with the one click installers.

Padraig Kane

could you do a video on onetrainer pls you are the best at explaining idk what it is but when you say something i just download it in my brain ye.... idk ahahha

HealingPaint

What is the total size needed on the C drive for this?? (since these always install giant models to the C:]users folder paths

GlobalGrooveWORLD

Hi! Problem with control net - **Node ID:** 278 - **Node Type:** WanVaceToVideo - **Exception Type:** RuntimeError - **Exception Message:** shape '[1, 132, 8, 90, 8]' is invalid for input of size 765360

E the G

Awesome stuff! What's the best way to inpaint with this model? Is it even possible?

Alex Cooper

Same. WanVanceToVideo shape '[1, 87, 8, 156, 8]' is invalid for input of size 876096

Dat Nguyen

has anyone tried the runpod install? i did everything exactly as instructed in the end of video. but it says a bunch of nodes are missing. this is even after i updated compfyui and restarted it.

Aitrepreneur

Redownload the installer, I just uploaded a more stable version

Aitrepreneur

Inpainting is possible, it works ok, I decided to take it out of the workflow (maybe I should have kept it maybe let me know if you want me to add it back) because it's better to just use the flux fill model, since it's the specialized inpainting model that works for pretty much anything

Aitrepreneur

Would be great to know if other people can test it and let me know if that works for them, I can link that in the post this way. I can't test anything apple related on my side so that's cool

Aitrepreneur

why specifically onetrainer? Showed recently AI toolkit that is easy to use and that you can use to train pretty much everything as well

Aitrepreneur

Come on man...watch the video...also if you know me and what I stand for, I wouldn't show it if you couldn't...

Aitrepreneur

No it's Wan, not flux. You need wan loras, you can find them all on civitai for example

Aitrepreneur

nope it's the portable version so you can have as many copies of comfyui as you want (I got 5 myself)

Number 6

Amazing work, Thank you! Is it possible to train Loras? I looked at AI-toolkit, but it obviously trains short videos rather than single images and would probably take a month to generate anything!

Aitrepreneur

You can actually train just on images, you don't need videos to train a Wan lora :)

Number 6

Your next video perhaps..? ;)

Raphael Paz

Is 5 minutes to generate 720p image with rtx 4080 normal? Using k_5

Padraig Kane

ok thanks do i need to upgrade i can only see the gigachad training nothing else so thats why i asked, where can i find the kohya ss trainer? or is there a way to train a human with your latest trainer?

Alpha Ghost 47

I have same issue with this - WanVaceToVideo shape '[1, 135, 8, 38, 8]' is invalid for input of size 330480

Nedo Braun

is this possible with ostris ai toolkit? the one you showed in your last video regarding Flux Kontext training.

Number 6

I'm actually trying that right now using the Ostris toolkit. I had to ensure only 1 frame was generated during the sampling phase or the job was going to take days, even on a Runpod H200 SXM!!

Virtamouse

this is really exciting. can you please give instructions for training character Lora for wan using the ai-toolkit? I tried and it seemed to just freeze up the tool when I selected wan2.1 14b as the model. although I've used your guide to train flux, sdxl, Kontext all without issues using ai-toolkit.

Jason Brashear

No Doubt. I love what you do. I just don't have a PC!

Elkin Marenco

After running WAN-IMAGE-MODELS-NODES_INSTALL.bat I get this "Traceback (most recent call last): File "F:\AI\ComfyUI_windows_portable2\ComfyUI_windows_portable\ComfyUI\main.py", line 138, in import execution File "F:\AI\ComfyUI_windows_portable2\ComfyUI_windows_portable\ComfyUI\execution.py", line 15, in import comfy.model_management File "F:\AI\ComfyUI_windows_portable2\ComfyUI_windows_portable\ComfyUI\comfy\model_management.py", line 221, in total_vram = get_total_memory(get_torch_device()) / (1024 * 1024) ^^^^^^^^^^^^^^^^^^ File "F:\AI\ComfyUI_windows_portable2\ComfyUI_windows_portable\ComfyUI\comfy\model_management.py", line 172, in get_torch_device return torch.device(torch.cuda.current_device()) ^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "F:\AI\ComfyUI_windows_portable2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\cuda\__init__.py", line 1026, in current_device _lazy_init() File "F:\AI\ComfyUI_windows_portable2\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\cuda\__init__.py", line 363, in _lazy_init raise AssertionError("Torch not compiled with CUDA enabled") AssertionError: Torch not compiled with CUDA enabled" I have a RTX 2060 with 12 GB of VRAM This is the info with mvidia-smi NVIDIA-SMI 571.96 Driver Version: 571.96 CUDA Version: 12.8 0 NVIDIA GeForce RTX 2060 WDDM | 00000000:09:00.0 Off | N/A | | 0% 44C P8 20W / 184W | 742MiB / 12288MiB | 15% Default | | | | N/A

Elkin Marenco

I reinstall ComfyUI and now everything it's OK

PTMarks

Do the new models work with the Img2Vid WAN or VACE workflows? I tried to use the new models but the video changed the details of the original image significantly. If I use the old models the video pretty much stays true to the original image.

Aitrepreneur

Seems a bit slow, have you tried with magcache activated?

Aitrepreneur

You can't, it's a different model you need wan loras

Aitrepreneur

I might do a video for Lora training for Wan

Aitrepreneur

No these are text to video model, you need the same but image to video and also reduce the number of steps if you use the fusionx model

Bob Bobenko

Really good model, thanks man!

E the G

Thanks for the reply! That would be fantastic if you could include it too. If not, no worries. I will check out the flux fill model in the meantime. Do you have any workflow recommendations for it?

Josh inmn

WOA why are you default installing into the Windows/system32 dir? I don't like that. What is your reasoning behind that? That's unsafe practice, running a bunch of scripts in your systems path folder.

Aitrepreneur

That's because you are running it with admin rights, don't do that, just run the installer normally

Josh inmn

Oh yeah, but also, Thank you! It's easy to see how much work you put into this stuff, plus I can just feel the frustration comments like mine can cause you, I'm sure. sry. I'm fairly certain I just couldn't deal with all of the Mes out there. I get going just readying them. So thanks for putting up with us. Anyway Easily worth the monthly patron support. You are enabling thousands who might never ever even try if they hadn't found your videos. So a genuine token of gratification for you my brother. 💪🏼

Josh inmn

OMG is that what happen? LOL my dude, sorry. I was trying some things in my $Profile last night and totally forgot. I can't stop the quest for ricing powershell terminal. Breaks everything, loads time is 1.5k ms, its aweful...but I love it. COOL good to know, thank you! hahahaha

Alson

i have an error saying Unsupported 160Bit application while it tried to run git-setup.exe as it is incompatible with my window version of 64-bit. what should i do?

Alson

ok it got resolved.

Alson

after the first installation where after installation it opened up automatically. But subsequently, after i closed the window, I was unable to find the file to open up the windows again. Please advise which file should i click to launch the UI

Kevin

I have the same issue. How did you resolve this, please?

Alson

you need to install git separately first

Padraig Kane

i trained i wan lora with no problems it should work but i did download 2 versions i downloaded a 2750 and 2500, with the q 5 model the 2500 lora work great and the 2700 didn't work, and with the q8 model the 2750 lora worked great and the 2500 one didn't at all

Cory Fitzgerald

I am getting the following error when I try to use Image to Image VACE control. I have tried researching this myself, and all the answers I found either didn't work, or I couldn't figure out. 025-07-15T19:10:22.731531 - got prompt 2025-07-15T19:10:23.008543 - gguf qtypes: F16 (693), Q5_K (489), F32 (149) 2025-07-15T19:10:23.080610 - model weight dtype torch.float16, manual cast: None 2025-07-15T19:10:23.081610 - model_type FLOW 2025-07-15T19:10:25.104112 - Requested to load WanTEModel 2025-07-15T19:10:26.700565 - loaded completely 9225.841395187377 5039.953125 True 2025-07-15T19:10:30.494252 - loaded partially 193.69487380981445 193.6948299407959 0 2025-07-15T19:10:30.717455 - 0 models unloaded. 2025-07-15T19:10:30.723460 - loaded partially 193.69483003616332 193.6948299407959 0 2025-07-15T19:10:30.915635 - !!! Exception during processing !!! shape '[1, 149, 8, 92, 8]' is invalid for input of size 887885 2025-07-15T19:10:30.917638 - Traceback (most recent call last): File "J:\Wan AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 361, in execute output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "J:\Wan AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 236, in get_output_data return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "J:\Wan AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 208, in _map_node_over_list process_inputs(input_dict, i) File "J:\Wan AI\ComfyUI_windows_portable\ComfyUI\execution.py", line 197, in process_inputs results.append(getattr(obj, func)(**inputs)) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ File "J:\Wan AI\ComfyUI_windows_portable\ComfyUI\comfy_extras\nodes_wan.py", line 258, in encode mask = mask.view(length, height_mask, vae_stride, width_mask, vae_stride) ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ RuntimeError: shape '[1, 149, 8, 92, 8]' is invalid for input of size 887885

Yu Yeh

We look forward to another great LoRa tutorial you always do for this model using ai-toolkit. I guess it is pretty much the same process but the way you explain is always helpful and fun.

NIGGRK

is there any way to run it on amd gpu?

Anders

Missing Node Types When loading the graph, the following node types were not found workflowFlorence workflowFlux Controlnet workflowNodes 3 workflowXLabs Sampler

Anders

after i pressed install missing node and rebooted, this still pops up

Franco Giova

you have to click the run_nvidia_gpu.bat file

Timmy

Wrong Array Size Error ~SOLVED~ I keep getting an error from the CLIPLoaderGGUF node, saying it "cannot reshape array of size ### into shape (###,####)" I get this error every once in a while with a new workflow, and I usually have to futz around with the model. However, in this case I have the exact same ones you have selected in the video (Used Patreon Downloader). Am I the only one who gets this? I was wondering if you could tell me what causes that error, and how to fix it? P.S. I have other Text to Vid workflows working. It is only some workflows, and I was hoping to understand the cause. I have a RTX4090. Unet Loader (GGUF): Wan2.1_T2V_14B_FusionX-Q8_0.guff Clip Loader (GUFF): umt5-xx-encoder-Q5_K_S.guff LoadVAE wan_2.1_vae.safetensors For the Lora I am only using: Wan21_T2V_14B_lightx2v_cfg_step_distill_lora_rank32.safetensors Since FusionX has the other Lora built in.

Timmy

So, a little more insight into this. I decided to ask Chat GPT what the issue was. In short it stated that: The incompatible piece is: umt5-xxl-encoder-Q5_K_S.gguf used as a CLIP loader -------------------------- umt5-xxl-encoder is not a CLIP model — it’s a T5-based language encoder. Specifically, it’s a text encoder from Google's mT5 (Multilingual T5) family, typically used for multilingual tasks, not vision-language alignment like CLIP. The CLIPLoader (GGUF) node is expecting a model trained in the CLIP architecture, which produces a specific latent shape (typically 768, 1024, or 1280 dimensions depending on the variant). When it tries to reshape output embeddings into (10240, 2816) (a CLIP-style tensor), it crashes — exactly the error you're getting. ----------------------------- But this appears to be the exact same clip loader that @Aitrepreneur uses in the video. I changed the clip model to: umt5_xxl_fp16.safetensors And all of a sudden it worked. I used Aitrepreneur's model downloader, and I think that "umt5-xx-encoder-Q5_K_S.guff" is the one that it downloaded. What am I missing here?

Timmy

Ok, I fixed it. Apparently I must have gotten a file with the exact same name from somewhere else that was not the correct file. I deleted the one that I had. I then manually went into aitrepreneur's model downloader bat file and used the link in there. Once I re-downloaded it, it started working right away. I think Hugging face has a file by "QuantStack" with the same name. I think that is where I got the one that was giving the issue. Not sure where aitrepreneur's came from. But take this with a grain of salt because I was really confused when I started all this a few days ago. So I am not positive where I got files, or if they were renamed. Either way, if you get the array size error, it is probably that you have a text encoder that is not made for "Clip" but one of the other formats. If all else fails, paste the output window in chat gpt, and tell it what nodes you have and their settings, and let ChatGPT figure it out lol. Hope this helps someone else.

Anders

haha stupid me, did not use any workflow so i downloaded the workflow and it works, Thank you :)

phail216

umt5_xxl_fp16.safetensors throws the same error for me

Aitrepreneur

That's why you don't ask Chatgpt, it doesn't know shit about comfyui. The reason why it didn't work isn't because "it's a text encoder not a clip" it's simply because the workflow is made to use the GGUF quantization model format and you are using a f16 safetensors format.

Sam

Any tips on keep the same character without training a lora?

Aitrepreneur

Flux Kontext is better suited for that then

Sea Tree

when i installed this and clicked i got the error unsupported 16-bit application

Wilhelm Zabłotnikov

Is it possible to train a LoRa model for Wan2.1 using photos? Or only with videos?

Sam

I believe you can just use photos

Sam

From what I see online anyway

Aitrepreneur

Which missing nodes? Have you tried first updating comfyui and reloading the workflow?

Wowza

How do I use it via Mac?

Sam

what for?

Alex

On runpod, I'm getting this error when restarting after updating ComfyUi: Traceback (most recent call last): File "/workspace/ComfyUI/main.py", line 136, in import comfy.utils File "/workspace/ComfyUI/comfy/utils.py", line 20, in import torch File "/workspace/ComfyUI/venv/lib/python3.11/site-packages/torch/__init__.py", line 409, in from torch._C import * # noqa: F403 ^^^^^^^^^^^^^^^^^^^^^^ ImportError: libcusparseLt.so.0: cannot open shared object file: No such file or directory

Aitrepreneur

I'm aware of the issue, I might have a fix but to get around this issue, just first, go inside comfyui and update all. Once everything is updated to the latest version, then you can actually use the installer, once everything is downloaded, go inside comfyui, manager and then click restart. Basically it's the same thing but instead of updating everything at the end, you do it at the beginning

Vyacheslav

Hello! :) Thanks for the workflow. It's really amazing. But I faced the same issue of the "image resize" node, when using VACE. You can simulate the bug by trying different images with different sizes.

Aitrepreneur

I've already updated the workflow to solve this error, just redownload it from the page

THE DEVIL

Okay, So I tried to install everything, but when i load up comfyui with run_nvidia_gpu.bat, and then i load WAN_IMAGE_GENERATION.json, I get this message below, saying that it's missing all kinds of node types, when it loads there are a lot of red nodes. Do you know how to fix this? er (rgthree) sse When loading the graph, the following node types were not found: • · · CLIPLoaderGGUF TrimVideoLatent WanVideoNAG • MagCache • workflowsampler and seed upscaler • UnetLoaderGGUF • workflowUPSCALERS • EmptyHunyuanLatentVideo • WanVaceToVideo • WanVideoImageResizeToClosest • workflowSampler 3 Nodes that have failed to load will show as red on the graph. Close по Text ter aut ed rat

THE DEVIL

I updated ComfyUI to the latest version through the manager but now run_nvidia_gpu.bat just loads a pure white window.

THE DEVIL

Reloading the window fixed it, now all it wants is three nodes: workflowsampler and seed upscaler workflowUPSCALERS workflowSampler 3 It says i can install through the manager, I'll try it out.

THE DEVIL

I wasn't able to get the remaining nodes installed, but it looks like it only affects two of the workflows, text to image should work. I'm not sure how it all works, but I'll try to figure it out.

THE DEVIL

No matter how i put the settings, i always get: UnetLoaderGGUF cannot reshape array of size 8396068 into shape (5120,5440) after loading for about 10 seconds. ------------------------------------------------------------- If i change Load VAE from anything else to wan_2.1_vae.safetensors I get: VAELoader Error while deserializing header: MetadataIncompleteBuffer File path: D:\DOWNLOADS 2\ComfyUI_windows_portable\ComfyUI\models\vae\wan_2.1_vae.safetensors The safetensors file is corrupt/incomplete. Check the file size and make sure you have copied/downloaded it correctly.

Aitrepreneur

can you dm me instead of all these comments, I can't follow up instead

Bob Bobenko

On Runpod It worked for a week, but now I am getting: ImportError: libcusparseLt.so.0: cannot open shared object file: No such file or directory After Update All and Restart server

Bob Bobenko

I've tried it on 3 different pods, with different GPU, and it just doesn't work. Looks like some libs updated. Try to fix your script please

Bob Bobenko

Fixed it! apt-get update apt-get install -y cuda-toolkit-12 libcusparselt0 And then restart an app with application manager on 8000 port

Testtn

You're the man Bob! Work for me last night then all of a sudden not working tonight... But your little trick there worked! Thank you!

Aitrepreneur

You don't have to do all this, just use the installer AFTER you updated comfyui, simple fix

Yves-Gaël Jacak

Hi ! Work like a charm ! But i have an issue, impossible to make picture look like art, like drawings, paintings...etc. What ever my prompts, it's always gonna be photo-like. Do i miss something ? Thanks :)

Aitrepreneur

yes, it's your prompts, ask chatgpt to make you a prompt for flux and ask what you want in details

myaiworld

i dont see WAN_IMAGE_GENERATION.json file in the attachments. Also runpod template link "aitrepreneur/comfyui" not provided int the article

Robert Elms

Hi. Successfully installed and ran on Runpod 3 days ago. When I run the WAN-IMAGE_AUTO_INSTALL-RUNPOD.sh now it appears to run successfully with 'Wan FusionX models and nodes are ready.' appearing in the terminal window. If I try to open Port3000 I get a message saying Port is not up yet. Any ideas?

Aitrepreneur

try to update comfyui first, then use the installer, comfyui has had an update that breaks a lot of things if the update is done AFTER the nodes are installed, so just do it before

Testtn

I did that the Tuesday, didn't work Wednesday, Bob's trick worked.

Cristo

Ive tried to install it and there's this error "7-Zip install failed ÔÇô install it manually then rerun this script."

Aitrepreneur

Uninstall 7 zip if there is an install already (it should be installed in the C:\Program Files\7-Zip folder ) then relaunch the installer or as it says reinstall 7 zip manually from the website

Yves-Gaël Jacak

I'll try that later today, thank you very much :)

Jonny

I am having the same issue. Did you manage to get it fixed?

Jack Maed

I'm getting an error when trying to run WAN ouputs. "Prompt outputs failed validation: UnetLoaderGGUF: - Value not in list: unet_name: 'Wan14BT2VFusionX-Q8_0.gguf' not in ['Wan14BT2VFusionX-Q5_K_S.gguf', 'Wan2.1_T2V_14B_FusionX_VACE-Q5_K_S.gguf', 'wan2.1-t2v-14b-Q5_K_S.gguf']"

VSEVOLOD KULAGIN

Hey, did you saw WAN NABLA finetunes on huggingface? Looks like it'll help improve both training and generation speeds.

Aitrepreneur

you need to chose the correct model in the model selection node (the first one)

Mr.Rie asd

It takes years to generate an image and it completely looks like FLUX...

Blend

When i run the comfy ui, it says its missing a node called "String Literal". There were originally more nodes missing but I managed to install all of them properly except for that one, and I don't know how to install it, or rather where to get it. I tried with a pack called "ComfyUI-Helper-Nodes" that at least I thought it included it but no luck.

Cristo

I didnt have it, now it’s intalled. Thanks!!

Aitrepreneur

can you send me a sceen of the parameters and error that says that

Ed Sullivan

I already have a manual install of ComfyUI installed on its own M2 Drive. I really don't want to install another version as the one I have now I've been using for a while and have it set up with SD#, Flux, etc. Do I just use the "WAN-IMAGE-MODELS-NODES_INSTALL.bat" installer for an existing installation. I want to check before i muck up my existing install or computer.

Stefan Djokic

I just install all of them inside confyui portable folder, never had a problem and didn't need a fresh one. P.S. this is Wan 2.1, new model has gone out 2 days ago, Wan 2.2 and it's much better, if you are able to set it up yourself, it's worth it

Cristo

I am a little bit confused. I had nothing installed on my computer. So I installed Git, Python and other stuff but not Comfyui. then I installed this 1-click installer for Wan. Was I supposed to have Comfyui firts then this? Or this is going to intall everything I need? Btw about the files Do I must to download and install Wan Image Model Nodes as well? I intend to use comfyui locally.

Robert Elms

I started a new pod and updated Comfyui before installing anything else, as suggested by Aitrepreneur. Now works perfectly

Paweł Pawłowski

RuntimeError: operator torchvision::nms does not exist

Blend

oh forget it, i was using the wrong workflow and just now realized, oops!

Orla

Running on RunPod dosent seem to work anymore. After hitting restart, it gets stuck on reconnecting... before telling me the port is not running and I have to start all over again and it just goes in a loop

Aitrepreneur

Update comfyui first then use the installer. Once it's done go back to comfyui in the manager and click restart then drag and drop the workflow inside

Orla

Thats what Ive been doing for weeks but today its not working anymore unfortunately

Aitrepreneur

DM me what you are doing your steps by steps

Orla

DM sent

Edgar Hoover

Hi, any update on this? same issue here.

Mark Mynhier

It says that I don't have 7 zip installed, but I do. Why, and what's the work around?

Matt

I use Runpod and it used to work fine until recently. Now I get the following error message: "Failed to get input node 0 for group node child 433:1 with slot 1".

Aitrepreneur

probably because it's not added to path. it should be installed in the C:\Program Files\7-Zip folder and added to path. But you can just uninstall 7-zip and relaunch the installer, it should reinstall 7zip correctly

Aitrepreneur

It happens for some reason sometimes, restarting and reloading the workflow usually fix the issue

Kevin Brunt

I just can't get it to work on RUNPOD. When I try to restart and refresh the comfy ui interface I get a message that the port is not yet up... I have been out of this for a few years and the deep end is deeper than I thought.

Holy Buckeye

i use runpod, everytime i drag the runpod install over in jupyter and hit the terminal, copy and paste the text, i get a ERROR, run this in your Comfyui root (models/ custom nodes/ present) and i am clueless to what the means

AP '

Please help me out, I use Runpod, for the installation of the runpod auto installer, everything worked well until I restarted my ComfyUI and it showed me "missing models, When loading the graph, the following models were not found" and then it gives me the safetensors file, what should i do ? I already tried to upload it in the workspace but it doesnt do anything

TrapHouseMag

Can you make a Runpod video? This is so hit or miss. Come on man!