WAN 2.2 💠 T2V / I2V / S2V / T2I🔹4-6 steps🔹Loop 🔹SVI Video extend 🔹WanVideoWrapper Workflow🔹K3NK
NOTICE: IF YOU ARE NEW TO COMFYUI, I DONT RECOMMEND STARTING WITH THIS WORKFLOW, IS AN ALL IN ONE WORKFLOW YOU WILL GET CONFUSED.
I UPDATED v2.3, YOU CAN CHOOSE IF YOU WANT RES4LYF OR WVW SIGMA SPLITTING
Since v1.8 im using subgraphs, you need latest comfyui with the latest frontend updates, also probably you will need to update Comfyui Manager.
For me was: setting nightly vers on the manager then update confy from the manager too.
* i allways leave the image overlay enabled, it will error if you dont have an image on it, sorry :S
v2.0
You can choose whether to merge with the loras or not, change the lora connection from the model loader to the set loras node, after that uncheck merge_loras
The toggled i added for radial sage, requieres Sage Attention to be installed, if you dont have it you should unplug the conexion i made to the attention type on the model loaders..
Be sure to update WanVideoWrapper nodes ..
The workflow works just fine with Wan2.1 if you bypass first sampler and model loader
This workflow is included with all my uploads as metadata, drag and drop a vid to comfyui.
I run a 4090 locally
You can use it as other modes if you activate this option for enabling the required nodes.
Make sure you load the correct models on the model loaders
Kijai's WanVideoWrapper Node
Wan2.2 Kijai's models
Lightx2v LoRa needed for lower steps (faster generations)
Wan22-Lightning Lightx2v alternative, newer and better in some scenarios
NSFW Clip Vision (not really needed in wan2.2)
RES4LYF Nodes
MMAUDIO
https://huggingface.co/Kijai/MMAudio_safetensors/tree/main
once installed replace nodes.py file in the mmaudio node folder with this one: https://github.com/phazei/ComfyUI-MMAudio/blob/master/nodes.py
WAN 2.2 💠 T2V / I2V
MMAUDIO NSFW
https://huggingface.co/phazei/NSFW_MMaudio (download and rename it to "mmaudio_nsfw_large_44k_v2_fp16.safetensors")
BOYONODES
K3NK IMAGE GRAB
Description
Added SVI with save latent node
SVI is usable now alongside the regular last_frame PNG sequence method, it will be using the .latent files for guiding SVI inference but the final video will be combining the PNG as allways
Added HuMo alongside S2V, for some cases S2V can handle the base image better, you will need to swap the conections of the subgraph to the oter in case you want to use HuMo or viceversa
I updated my K3NK image grab node to work with the .latent files, you need to update the node
FAQ
Comments (20)
Great LoRAs, but your workflow gives unreasonable anxiety. Not recommended for claustrophobes...
I wish there was an AI way to reorg into subgraphs.
sadly subgraphs doesnt support group bypasser/muter ve already tried
Claustrophobia has nothing to do with clutter. It has to do with enclosed spaces.
As soon as I opened this, i was like wtf is this dependency hell?! the amount of custom nodes that need to be installed gives me anxiety. Can you find a way to run this natively, without forcing people to spend days debugging this mess?
please tell me what that list was so i can at least know.. thanks
Ill agree, opening it is a dependecy hell. But think about what you get, an workflow that is carefully crafted to achive results that "native" workflows will not, were you can customize and get alot of different results. My advice start with simpler workflows and work yourself up to more complex. @K3NK Thank you for contributing, i use ur workflow to get inspired when creating my own, its a great way to learn. keep up the good work!
@kaysismail433 thanks for the kind words mate, i know noodles are not liked, but i rather follow a noodle than start seeking for a get/set node that i cant find.. xD
i might remove all the "extras" im not using and try to subgraph the shit out of it.. but the last time i tried doing that i gave up
working perfect
seems to me i cant install the missing nodes? any way around this?
please share the nodes you cant install. i adedd most of them in the description
Hey there is some problem with the nodes i cloned every single one and after that i have problem with this nodes:
-FramePackFindNearestBucket
-DownloadAndLoad
-GIMMVFIModelGIMMVFI_interpolate
-Sigmas Split Value
After that there is announcement all missing nodes instaled and when i want run it i2v there is problem ,,Node ID '#254''
Nice one!! But please can someone explai nme how to use and install Torch Compile , i'm only missing that ;(
There is a long list of dependencies and versions you need to have to be able to use it. If you're using the wrapper you should be running the latest version of comfy and have everything up to date. It's not just a matter of installing a node. Try python -m pip install everything. And update printer drivers. Inkjet, not laser. Better yet, do a fresh install with the Tavris installer. It's really well written. If your computer can run what is required, it will get you up to date without you having to know what you are doing. Which includes everyone, at some point. All of that said, don't expect something magical from the compiler. It might not do what you think it does. Unless you like recompiling every time you change anything. Which is what it does. Just so you know.
I was just about to upload a 4 stage SVI wf. Stupidly I put in 16 4-prompt sets and now I have to fill them all in. I'm going to steal some of yours. It's always interesting to check out different implementations of SVI embeds, especially when it's someone who actually knows what they're doing. Ahhh, damnit string to combo, that's it. Was trying to remember how to switch the LoRAs along with the prompts. Duh. Combo data is evil. Your commitment to efficiency is humbling. Tracing execution in your workflows is like a textbook for comfy logic. Beautiful. I hope people are paying attention. You've got so many great ideas in here. Two latent thumbs up.
Hi, could you please provide a workflow just for SVI? I'm finding it very difficult to download everything.
Thanks for Wonderful workflow
It seems like Ollama is required for this workflow somehow? But it's not clear how to run it in this setup.
(Solution) I had difficulty installing ComfyUI-WanVideoWrapper. Originally ComfyUIManager said that there's 146 nodes and 34 conflicts in this workflow, but I seemed to have it installed and and it didn't work. So I directly git cloned into the existing installation folder and overwrote any existing files, restarted the server (through Stability Matrix), and presto! No more error.
Great workflow as always. Is there anyway to edit the MMaudio output save location. Was looking around the WF, but can't quite figure out if it is possible.
Nice work! seems overwhelming but works well! Will you make NVFP4 Support version?