Nvidia just unveiled its new AI production model, dubbed Latte3D, during GTC 2024. Latte3D appears to be ChatGPT on extreme steroids. I’m a 3D text model that takes simple, short text prompts and turns them into 3D objects and animals within a second. Much faster than its older counterparts, Latte3D works like a virtual 3D printer that could be useful for creators in many industries.
Latte3D was created to simplify the creation of 3D models for many types of creators, such as those working in video games, design, marketing, or even machine learning and robotics training. In Nvidia’s model demo, it looks extremely simple to use. After a quick text prompt, the AI creates a 3D model and shortly afterwards completes it in much more detail. While the end result isn’t as lifelike as OpenAI’s Sora, it’s not meant to be – this is a way to speed up the creation of assets instead of having to create them from scratch.
The model creates several different options for the user to choose from, and Nvidia says these shapes can be “optimized for higher quality in minutes.” The designs can then be exported to different platforms, such as Nvidia’s Omniverse, and can be modified to match the desired end result. Nvidia trained Latte3D using its Ada A100 Tensor Core GPUs and supported the training with ChatGPT prompts to prepare it for interaction with real users.
Get our weekly breakdown of the technology behind PC gaming
Currently, Latte3D can only create objects and animals. To that end, it seems to do a solid job of distinguishing between different animals, textures, and types of objects. Nvidia demonstrated these capabilities by presenting objects such as an amigurumi (crochet) common crane or an origami sphynx cat. The model was taught to recognize different species, so it can tell the difference between an Italian Greyhound and a Shiba Inu.
Creators who want to use Latte3D to do more can train it on a different dataset, be it plants or household objects, and later use it for their own purposes. Nvidia presents some interesting use cases here, such as training personal assistant robots before they are deployed. It’s easy to imagine that Latte3D will be useful for game developers, but the possibilities go far beyond gaming scenarios.
Sanja Fidler, vice president of AI research at Nvidia, noted how much faster Latte3D is compared to its predecessors: “A year ago, it took an hour for AI models to generate 3D graphics of this quality — and the current state of the art is now about 10 to 12 seconds. We can now produce results an order of magnitude faster,” said Fidler.
The recent announcements about the use of artificial intelligence in game development are all quite groundbreaking, and Nvidia’s Latte3D adds to a growing list of tools that may one day completely change the process of creating a game. For example, Nvidia just recently unveiled non-player characters (NPCs) with dialogue generated entirely by AI. Meanwhile, the latest Unreal Engine update can create movie-quality graphics in games in real-time, all with the help of machine learning.
Editors’ recommendations