Text-to-Motion Model

A multimodal generative model that generates human skeletal motion sequences from natural language descriptions, outputting time series of joint rotations or displacements, typically relying on parametric human body models such as SMPL.