NVIDIA introduces the LATTE3D model: artificial intelligence for creating a text-to-3D model

NVIDIA introduced LATTE3D, a new generative AI that can convert text commands into 3D models, at GTC. Within one second, LATTE3D creates photorealistic 3D objects from natural language input – ready for use in game development, design, and robotics.

Unlike existing text-to-3D systems like MVDream or 3DTopia, LATTE3D delivers similar or higher quality in just 400 milliseconds. This speed advantage promises huge workflow improvements in heavy 3D industries. You can initially focus on speed rather than quality and later create an optimized version instead of having predefined settings. To date, LATTE3D has been trained on everyday objects and animals, but with new datasets the model can be expanded to include additional categories.

Lightning-fast 3D creation from text descriptions can improve many workflows. Instead of searching through libraries of 3D assets or designing objects manually, a simple text input will soon be enough to import any 3D content into rendering applications.

Although LATTE3D is still in the research phase, NVIDIA is already planning to expand it to create 4D animation from text. At the moment there is no announcement of a commercial product launch. Details about the work can be found here Directly from Nvidia.

14 AI Generators for 3D Models – 2024 Overview