DC-Gen applied to next gen video generation models?

I feel like the video gen models are held back only because of the inference time limitaions. A DC-Gen Wan2.2 T2V and I2V or the upcoming Wan2.5 series models (if they open source) would be fantastic.

Edit: NVM, I saw you have another project for video models.I hope you guys would do the same to any upcoming next gen open source video models, kind of like the nunchaku devs.