Generate a video based on a text prompt using Mochi
Apply the motion of a video on a portrait
Inference with 4/8-bit quantization