OpenAI has done it again with its latest innovation, the text-to-3D model known as Shap-E. Shap-E is a generative AI tool that turns text prompts into three-dimensional objects with fine-grained textures and complex shapes.
OpenAI has used a technology known as neural radiance fields (NeRFs) to develop Shap-E, which is the same technology used in virtual and augmented reality to create photorealistic environments. Shap-E generates more accurate responses than a similar tool called Point-E, introduced by OpenAI in 2022.
Shap-E applies NeRFs to common diffusion models to create the shape and texture of an object from text prompts, making the process significantly faster than Point-E. The tool is currently available for free on GitHub and runs locally on a PC. However, setting up the tool and getting it to run is a mammoth task, and the output is currently passable.
The potential of Shap-E is vast, with examples demonstrated by OpenAI that include a bowl of food, a penguin, and a campfire, all generated within seconds. The tool can benefit movies, virtual reality experiences, engineering simulations, scientific research, and other applications. Although the output from Shap-E is low-res, with key details missing, the tool is touted to be revolutionary in 3D modelling and rendering.
OpenAI has been on a roll with its innovations, starting with the revolutionary ChatGPT and its subsequent version with enhanced capabilities. Shap-E is yet another breakthrough from OpenAI after the success of its text-to-image generation tool Dall-E 2. With Shap-E, OpenAI has leaped forward in the field of generative AI tools and has shown yet again why it’s considered a powerhouse in the AI domain.