⚠️This workflow is experimental and blends both T2V and I2V in the same workflow, you CANNOT control the input image for I2V ⚠️
If you're looking for dedicated T2V/I2V workflows with LoRA support, visit my profile.
This is a workflow I made for video extension with Wan with support for different LoRAs for each pass.
For example you can generate the initial video using a Film Noir LoRA and the I2V extension with an undress LoRA to create a scene where a woman enters the house in a film noir style and then undresses.
Description
FAQ
Comments (14)
so basically you can make small movies with this? if i add the same prompt and same loras would it make a same style video that is longer than 4 seconds? would be amazing
finish the I2V process please, you set only T2V
Could you elaborate please?
@Hearmeman the group node for I2V not working
@dominic1336756 Can you send me a dm?
I just downloaded the workflow from civit and have no problem running it
@Hearmeman Then where is the image input node for I2V for exemple??
@dominic1336756 You did not understand the purpose of this workflow.
You generate the initial video with T2V, it then automatically extends with I2V and you can use a dedicated prompt and different LoRAs for the I2V process.
@Hearmeman Thanks for sharing your workflow it works great! However, for me (and about 90% of the people I know), it’s completely useless since we all use i2v. Your workflow forces us to use t2v first and then proceed with i2v. Couldn’t you just add switches to enable or disable the t2i part of your workflow? It would be great if we could simply deactivate the t2v section, upload an image, and continue with the i2v process using LoRAs. Anyway, thank you for your time, effort, and, of course, for sharing your work!
@arman947 You are more than welcome to look at my profile and use my dedicated I2V/T2V/V2V workflows.
This shouldn't be your main Wan workflow and this is more for fun and experimentation.
I'll change the description of the workflow so it's clearer.
@Hearmeman Ok then..Sorry i m looking for I2V workflow...
@Hearmeman Thank you !!!!
Thank you! I was looking for something like this to experiment with. I only just started with ComfyUI and was not able to put this together myself. <3
The Extension Part is unfortunatelly useless, not an issue with this Workflow but a problem with the Models not able to understand how to continue / exten the Image. Played around with this but the extended Part just did whatever it wanted, instead of following the prompt.