You can now run one of the best AI animation models on CPU with no VRAM at all, or on ultra-low VRAM cards.
It’s NVIDIA’s Kimodo — the new C++ port (kimodo.cpp) makes it fully local and much more accessible. No heavy Python + CUDA stack required.
Repo: https://github.com/localai-org/kimodo.cpp
In this video I show how to install it, which skeletons actually work (Soma / G1), native GLB export, retargeting to Mixamo, and the path into Unreal Engine.
—
It’s NVIDIA’s Kimodo — the new C++ port (kimodo.cpp) makes it fully local and much more accessible. No heavy Python + CUDA stack required.
Repo: https://github.com/localai-org/kimodo.cpp
In this video I show how to install it, which skeletons actually work (Soma / G1), native GLB export, retargeting to Mixamo, and the path into Unreal Engine.
—
- Категория
- Бесплатные курсы
Комментариев нет.
