Robot's "spatial sense" ushered in a new solution: Ant Lingbo's new visual model, 1.1 billion parameters to challenge 7 billion-level capabilities

Robot Foresight (Public account: robot_pro) Author | Li Shuiqing Editor | Moying Robot Foresight reported on July 7 that today, Ant Lingbo dropped two "bombs" at the same time: it officially released the spatial perception model LingBot-Depth 2.0, and simultaneously open sourced its self-developed visual base model LingBot-Vision. A complete set of technical solutions from the bottom layer of robot visual representation to depth perception capabilities was unveiled for the first time. ▲Posted by Ant Lingbo (Source: Social Media What supports this achievement is LingBot-Vision, the “space-native” visual basic model released for the first time by Ant Lingbo. Once released, these two models attracted a lot of attention on the Internet. Rohan Paul, a well-known technology self-media, posted on the social platform Domestic 3D vision leader Aobi Zhongguang has integrated LingBot-Depth 2.0 into data collection equipment and SDK, and plans to launch an integrated camera integrating the commercial version of LingBot-Depth by the end of the year. GitHub address: https://github.com/Robbyant/lingbot-vision Hugging Face address: https://huggingface.co/collections/robbyant/lingbot-vision ModelScope