Nvidia released the CALM AI model: training virtual characters to simulate 5 billion human movements

CTOnews.com August 11 news, NVIDIA recently partnered with Israel Institute of Technology, Baylan University and Simon Fraser University to publish a technical paper on CALM AI model.

Nvidia stands for Conditional Adversarial Latent Models, which are used to train custom virtual characters.

Nvidia says 10 days of training in the real world is equivalent to 10 years of training in the simulated world.

After training, CALM AI models can simulate 5 billion human movements, including walking, standing, sitting, running, fighting with swords and other human movements.

Nvidia says CALM captures the complexity and diversity of human motion and is able to directly control the motion of characters.

The results show that CALM learns semantic motor representations, enabling control over the generated motor and style regulation at higher levels of task training. After further training, users can control these characters with interfaces similar to those found in video games.

CTOnews.com attaches Nvidia CALM official introduction and GitHub page, interested users can read in depth.