At the GTC2025 Global Conference, Qunhe Technology, a prominent member of Hangzhou's "Six Little Dragons," unveiled its open-source spatial understanding model, SpatialLM. Built upon a robust large language model, SpatialLM constructs a sophisticated 3D scene semantic generation framework. Leveraging video input, this model adeptly identifies and comprehends spatial point cloud data, performs intricate reasoning, describes the scene, and ultimately outputs the information in a natural language format. This groundbreaking innovation underscores Qunhe Technology's formidable expertise in the realm of 3D technology.
