Fetching the paper…

Focusing on What Matters: Object-Agent-centric Tokenization for Vision Language Action models · Around