The workflow im utilizing since wan2.1 using Kijai's WanVideoWrapper nodes, with latest nodes versions updates works like a charm for wan2.2, the workflow itself i saw inside Memorable's videos. I dont know if his the author, all th credits to him nonetheless 😎, i was using the "second pass" since i put my hands on that workflow 😆so for me using the double sampler is no biggie.
For this "version" of the workflow i just removed what i was not using and moved things arround to accomodate the double model/lora stack stuff.
This workflow is included with all my uploads as metadata, drag and drop a vid to comfyui.
I run a 4090 locally
Â
You can use it as I2V if you activate this option for enabling the required nodes.
Make sure you load the I2V models on the model loaders
Kijai's WanVideoWrapper Node
Wan2.2 Scaled models
Lightx2v LoRa needed for lower steps (faster generations)
Wan22-Lightning Lightx2v alternative, newer and better in some scenarios
Added looping capabilites for I2V
Ungrouped the nodes i grouped, it sucks, you cant group nodes just like that, set/get nodes wont work if grouped, and you end up with a big ass node ..
Added color matching for better looping with last/first frame
Re-arranged overlay image thing so it wont interfere with the color matching
Toggling sections will now enable-disable the torch compile and other nodes like florence for prompts in i2v
Firmgrain can be an issue with long videos, thats why is outside the group, just in case you want to enable it, seems like is better for T2V than I2V
Go ahead and upload yours!
Your query returned no results – please try removing some filters or trying a different term.