Diffusers Implementation

[linjing7]

Hi, thanks for your great work! I have successfully run your code and got impressive results. Congratulations!

I'm wondering if you have implemented your work with diffusers instead of SAT. It seems that most mainstream video generation models (CogVideo-X, HunYuan-Video) are primarily implemented with diffusers.

[jshilong]

Thank you for your recognization! I was really worried before the code release, as I was not sure whether everyone would be satisfied with the results. Your positive feedback has given me immense confidence.

Regarding the integration with diffusers, I don't have much experience in it. If the load of such task is not heavy, I'll strive to add this feature as soon as possible.

[nitinmukesh]

@jshilong

Yes please integrate in Diffusers. It will be helpful for many

[linjing7]

Hi, here is a link for your reference. I think it would not take lots of time for conversion since FlashVideo adopts a similar architecture as CogVideoX.

[ghunkins]

I opened an issue on Diffusers to track this! Very excited for this, thank you @jshilong .

huggingface/diffusers#10767