Gondola: Grounded Vision Language Planning for Robotic Manipulation
A modular manipulation framework that separates high-level planning from low-level control and coupling grounded plan generation with a 3D-based execution policy, this framework achieves state-of-the-art performance on the challenging GemBench benchmark and demonstrates promising transfer to real robots.
Shi-Zhe Chen, Ricardo Garcia, Paul Pacaud et al.
· 2 citations