With a new artificial intelligence model, NVIDIA hopes to take the sting out of creating virtual 3D worlds. According to NVIDIA, GET3D can generate characters, buildings, vehicles, and other types of 3D objects.
The model should also be able to create shapes quickly. According to the company, GET3D can generate around 20 objects per second using a single GPU.
The model was trained using synthetic 2D images of 3D shapes taken from various angles. NVIDIA claims that feeding around 1 million images into GET3D using A100 Tensor Core GPUs took just two days.
MORE FROM RAVZGADGET: TikTok Has Rolled Out Comment ‘Dislike’ Button Globally
The model can create objects with “high-fidelity textures and complex geometric details,” NVIDIA’s Isha Salian wrote in a blog post. The shapes GET3D makes “are in the form of a triangle mesh, like a papier-mâché model, covered with a textured material,” Salian added.
Because GET3D generates objects in compatible formats, users should be able to quickly import them into game engines, 3D modelers, and film renderers for editing.
That means developers may find it much easier to create dense virtual worlds for games and the metaverse. Other use cases mentioned by NVIDIA include robotics and architecture.
According to the company, GET3D was able to generate sedans, trucks, race cars, and vans based on a training dataset of car images. After being trained on animal images, it can also produce foxes, rhinos, horses, and bears.
NVIDIA notes that the larger and more diverse the training set fed into GET3D, the “more varied and detailed the output.”
Because GET3D generates objects in compatible formats, users should be able to quickly import them into game engines, 3D modelers, and film renderers for editing.
MORE FROM RAVZGADGET: Twitch Said It Is Banning Major Gambling Websites In October
That means developers may find it much easier to create dense virtual worlds for games and the metaverse. Other use cases mentioned by NVIDIA include robotics and architecture.
According to the company, GET3D was able to generate sedans, trucks, race cars, and vans based on a training dataset of car images.
After being trained on animal images, it can also produce foxes, rhinos, horses, and bears. NVIDIA notes that the larger and more diverse the training set fed into GET3D, the “more varied and detailed the output.”
