Join this webinar to learn how to curate state-of-the-art 3D deep learning architectures for research.

Research in 3D computer vision and AI have been on the rise, but researchers have lacked good utilities to make 3D models ready for deep learning. To bridge the divide, NVIDIA released Kaolin to accelerate 3D deep learning research and move 3D models into the realm of neural networks.

Kaolin is a PyTorch library for accelerating 3D deep learning research with efficient implementations of differentiable 3D modules for use in deep learning systems. It also mitigates the need to write wasteful boilerplate code while packaging together several differentiable graphics modules, including rendering, lighting, shading, and view warping. The functionality loads and preprocesses popular 3D datasets and native functions to manipulate meshes, point clouds, signed distance, and voxel grids. Kaolin supports an array of functions and metrics for seamless evaluation and visualization to render the 3D results.

• How researchers use Kaolin to accelerate 3D deep learning research

• An overview of 3D deep learning tasks, such as 3D classification, 3D segmentation, single-image 3D reconstruction, and differentiable rendering

• How to train a neural network for a custom task, like avoiding collisions with JetBot 

• Applications and challenges of using Kaolin to curate state-of-the-art 3D renderings

Content

content goes here