This is a very simple ComfyUI beginner friendly Image-to-Video workflow that will work for a single input image with a single WAN 2.2 Rapid All-in-One GGUF model with multiple LORAs (This workflow does require ComfyUI's LORA Manager plugin to function. This is further improvement of my older WAN 2.1 Image-to-Video workflow. This uses a very optimized WAN 2.2 Rapid All-in-One model's quantized (GGUF) version to keep the workflow simple and fast so relatively affordable GPU's (8GB / 12GB Nvidia or Radeon GPUs or some more affordable Apple Sillicone Macs) can run WAN 2.2 with better system resources efficiencies without the burden of loading two seperate models for LOW & High Noise processing, the AIO model handles both. It is a good idea to install both ComfyUI manager and LORA Manager plugin to help you easily download and manage Checkpoints, LORAs and other resources. Not only these two are helpful for this workflow but they will help you a lot in any other cases). You need a Hugging Face account to download your necessary WAN 2.2 Rapid All-in-One files (Details are mentioned below). Make sure you install GGUF addon for ComfyUI using ComfyUI manager and place the correct files in correct places. Also check out my other workflows for SD 1.5 + SDXL 1.0, WAN 2.1, Chroma, QWEN, HiDream and Flux.
How to use this -
#1. First select your input image for Image-to-Video generation, then
#2. just select your desired WAN 2.2 Rapid All-in-One GGUF model first and after this load
#3. one or multiple matching LORA(s) from LORA Manager to start
#4. then input your positive and negative prompts.
#5. select how many videos you want (Change the number besides the "Run" button)
#6. select video resolution, CFG, steps etc. settings
#7. finally press the run button to generate. That's it..
*** This workflow will also function if you know how to properly disable the LORA Node and bypass it. Also with some modifications this will work with WAN 2.2 5b GGUF models
Enjoy!
### To use this workflow you need to log into huggingface and download necessary files from there (I also included a text file on the archive that has the workflow file, in which you will find even more links for essential downloads for my other workflows) -
## WAN 2.2 Models
===============================================================================================================
### Download Links for WAN 2.2 Checkpoints
https://huggingface.co/QuantStack/Wan2.2-TI2V-5B-GGUF/resolve/main/Wan2.2-TI2V-5B-Q3_K_M.gguf
### Download Links for WAN 2.2 Encoders
### Download Links for WAN 2.1 VAE (WAN 2.2 14b models use WAN 2.1 VAE)
### Download Links for WAN 2.2 VAE (Only for WAN 2.2 TI2V 5b model, WAN 2.2 14b models use WAN 2.1 VAE)
### Download Links for WAN 2.2 Image-to-Video Lightning Lora
https://civarchive.com/models/1585622?modelVersionId=2337903
Description
FAQ
Comments (13)
Thanks for your workflow, it works pretty well and fast
I have a question, if I want to use a Lora, for example, WAN General NSFW model is that possible? Because that model has 2 loras, high and low
You can use any LORA you want or No LORAs at all, that's how I built this workflow.
@sarcastictofu I see, in the case of High/Low Loras, should I use both? or only one?
@cyberyisus You can use one (high or low), or both.. I usually use LOW because these are kinda also compatible with WAN 2.1 (I do use WAN 2.1 sometimes as they may produce more Artistic results, 2.2 for me is good at more poses and anatomy).. also check my WAN 2.1 LORAs. see my profile.
@sarcastictofu Legend
Thanks
Where are the 2 loras from the manager?
It's upto you what LORAs you want to keep, it will work with zero LORAs, just one LORA or multiple LORAs (whichever number of LORAs you wish to have).. I have seen workflows that are fixated on LORAs.. sometimes they are needed (like 4 or 8 steps LORA or specific task based one but in this workflow I left it open for users to figure it out.. opted for ultimate freedom as I explained in the description and Readme file.
If you exactly need those two here are they - Lightning LORA - https://civitai.com/models/1585622?modelVersionId=2090344 WAN Nipple Fix LORA - https://civitai.com/models/1905501?modelVersionId=2156804
Doesn't seem to use my input image it just makes something original. Any ideas?
Maybe you downloaded the T2I model
Take a look at the screenshot I provided and compare it with your current settings, maybe you got T2I model or used too high or low values of something..
Yeah that was it, was using T2V thanks!
Dear author, thank you very much! a lot of things work out, but not everything! is it possible to generate sex on this wan2.2 model? I'm having trouble with penetration, and instead, the girl gets a "device"))) Maybe I need auxiliary LORA for this?
And maybe someone can tell me how to do test generations, short ones, it takes a long time to create a video on my 4060 rtx, and the desired action doesn't start in time.
Thank you in advance to everyone who can provide some suggestions.