https://research.nvidia.com/labs/sil/projects/kimodo/
Kimodo is a kinematic motion diffusion model trained on large-scale optical mocap data. It is controlled through text and constraints to generate high-quality 3D human and robot motions.
It seems to run on 17GB VRAM, but I hope this can be done on less, like 7 GB for example ?
Alternative make it work in a cloud as a new edition to the character creator ?
Add metahuman format to character creator ?
Etc.