Train Physical AI With Real-World Simulations Using fVDB.

For physical AI to operate in 3D spaces, it must have spatial intelligence—the ability to understand and reason in the real world. Achieving spatial intelligence involves converting the world around us into AI-ready virtual representations that the model can understand. fVDB, developed by NVIDIA, is an open-source deep learning framework for sparse, large-scale, high-performance spatial intelligence. It builds NVIDIA-accelerated AI operators on top of OpenVDB to enable reality-scale simulations for training physical AI.

Read the full blog: https://blogs.nvidia.com/blog/fvdb-bigger-digital-models/

Read the fVDB technical blog: https://developer.nvidia.com/blog/building-spatial-intelligence-from-real-world-3d-data-using-deep-learning-framework-fvdb/

Watch: Introducing fVDB: Deep Learning Framework for Generative Physical AI with Spatial Intelligence https://www.youtube.com/watch?v=6JCp0hXb3S4

Apply for the fVDB Early Access Program: https://developer.nvidia.com/fvdb/early-access-form

Check out the OpenVDB Github: https://github.com/AcademySoftwareFoundation/openvdb

Learn more about fVDB: developer.nvidia.com/fvdb