ChatGPT creator OpenAI has released Point-E, a new AI tool to generate 3D images in minutes. Point-E, unlike other 3D imagers, does not require a high-end computer to work.
What is ChatGPT?
Chat Generative Pre-trained Transformer is a chatbot developed by OpenAI. ChatGPT is based on OpenAI’s GPT-3.5 Large Language Model and works well with both supervised and reinforcement learning techniques. ChatGPT and GPT 3.5 were trained on the Azure AI supercomputing infrastructure. ChatGPT can chat, answer questions, create content, write and debug code, test, manipulate data, explain and tutor, and more!
What is ChatGPT used for?
ChatGPT can provide information in clear and simple sentences instead of just a list of internet links. You can explain concepts in a way that people can easily understand. You can even generate ideas from scratch, including business strategies, Christmas gift suggestions, blog topics, and vacation plans.
ChatGPT creator OpenAI has released Point-E, a new AI tool to generate 3D images in minutes. Point-E, unlike other 3D imagers, does not require a high-end computer to run and can generate a model in less than two minutes with a single Nvidia V100.
OpenAI, the company behind the AI-powered ChatGPT chatbot and Doll-e text-to-image generator, has released a new tool that can generate 3D objects based on simple text input. Dubbed Point-e, its open source is available on Github, though it’s a bit tricky to test as users will need to be well-versed with command line tools, and the system requires Python, unlike ChatGPT, where users can register. to a website and test its capabilities.
The Point-E developers have also published a research paper explaining how the platform works and what its limitations are. They claim that the Point-E, unlike other 3D imagers, does not require a high-end computer to run and can generate a model in less than two minutes with a single Nvidia V100 GPU.
How does Point-e work?
In a nutshell, similar to OpenAI’s Dull-E, Point-E can generate 3D models with simple English commands. The document shows some bizarre examples like “a corgi wearing a red Santa hat,” “a multicolored rainbow pumpkin,” “a pair of 3D glasses,” and “an avocado chair, a chair that mimics an avocado.” Although the tool does not produce a 3D model in the traditional sense, it does produce a series of data points that represent a 3D shape. The tool processes the final output after parsing the input based on the “several million 3D models” it has already parsed.
wrote in the article titled “Point E: A System for Generating 3D Point Clouds from Complex Signals.” To create a 3D object from a text indicator, we first sample and then sample an image using a text-to-image model. A 3D object is optimized in the sample image. Both steps can be performed in several seconds and do not require expensive optimization procedures. According to the research work, Point-E is capable of efficiently producing diverse and complex 3D shapes conditioned by a text message. Our approach can serve as a starting point for future work in the field of text-to-3D synthesis. Interestingly, the Point-E researchers used OpenAI’s ChatGPT to write the paper.
The developers claim that the 3D objects generated by Point-E can help a lot in a wide range of applications such as virtual reality, gaming, and industrial design.
E-Point Limitations
Similar to the Dall-E 2D imager, Point-E also does not parse the input and its final output appears in low resolution. Also, the end result does not capture the “microscopic shape or texture”. But the Point-E technique can be improved as it analyzes more images from the real world. Once the system is improved, it can effectively challenge Google’s Dream Fusion, which produces more accurate results but requires powerful hardware.
-Animesh Sharma