Closed Bensong0506 closed 3 weeks ago
I see, thanks. Is there any technical difficulty supporting Opensora related models, as they are both 3d VAEs similar to Cogvideo models, I think it should be very similar? Thanks a lot for your reply!
I believe there is no technical difficulty. xDiT has already supported Latte. As far as I know, OpenSoras build their model architecture based on Latte? I am not sure. If you would like to apply xDiT on OpenSoras, we can help you on this topic.
Thanks a lot, I will first try the CogVideoX and see if everything works.
We found that there are two open-soras in China. Both of them are not interpreted in huggingface diffusers. We are not sure if they really be applied by users worldwide. We will support them if they are in the diffusers.