It's not remixing training data. It is actually generating video using a diffusion transformer architecture. More info here: https://www.reddit.com/r/LocalLLaM

It's not remixing training data. It is actually generating video using a diffusion transformer architecture. More info here: https://www.reddit.com/r/LocalLLaMA/comments/1aspxox/explanation_of_openais_sora_by_jim_fan_nvidias/ https://www.reddit.com/r/LocalLLaMA/comments/1aspxox/explana...