This article proposes the ui-r1 framework that extends the prediction tasks of reinforcement action based on rules to gui
The supervised fine adjustment (SFT) is the standard training paradigm for large language models (LLM) and graphic user interface agents ...