Stability AI and Tripo AI join forces to release TripoSR, a game-changing technology that generates high-quality 3D models from single images in under a second.
In a groundbreaking development, Stability AI and Tripo AI have partnered to introduce TripoSR, a cutting-edge 3D object reconstruction model capable of generating high-quality 3D models from a single image in less than a second. This innovative technology is set to revolutionize various industries, including entertainment, gaming, industrial design, and architecture, by providing a fast and accessible solution for visualizing detailed 3D objects.
Key Takeaways:
Performance and Accessibility
TripoSR outperforms other open image-to-3D models, such as OpenLRM, by generating draft-quality 3D outputs (textured meshes) in around 0.5 seconds when tested on an Nvidia A100. Not only is TripoSR faster, but it is also more accessible, as it can run efficiently even without a GPU. This accessibility makes it an ideal solution for a wide range of users and applications.
Technical Improvements and Data Preparation
The development of TripoSR involved several technical improvements over the base LRM model, including channel number optimization, mask supervision, and a more efficient crop rendering strategy. The training data preparation incorporated diverse data rendering techniques to better represent real-world images, significantly improving the model’s ability to generalize. A CC-BY, higher-quality subset of the Objaverse dataset, was carefully curated for the training data. For more details on the technical aspects, refer to the technical report.
An Invitation to Explore and Contribute
Stability AI and Tripo AI invite developers, designers, and creators to explore TripoSR’s capabilities, contribute to its evolution, and discover its potential to transform their work and industries. The code for the TripoSR model is now available on Tripo AI’s GitHub, and the model weights are available on Hugging Face.
You can read the technical report for more details.
Source: Stability AI
Grow your business with AI. Be an AI expert at your company in 5 mins per week with this Free AI Newsletter
In February 2024, OpenAI introduced Sora, a video-generation model capable of creating one-minute-long, high-definition videos.…
Alibaba Group Holding has unveiled Qwen2, the latest iteration of its open-source AI models, claiming…
Google has rolled out a major update to its AI-powered research and writing assistant, NotebookLM,…
Stability AI, renowned for its revolutionary AI-powered art generator Stable Diffusion, now unveils a game-changing…
ElevenLabs has unveiled its latest innovation: an AI tool capable of generating sound effects, short…
DuckDuckGo has introduced a revolutionary platform enabling users to engage with popular AI chatbots while…