RMTL: Reinforced Micro-task Learning for Long-Horizon Manipulation with VLM Rewards figure
AlphaXiv 中文概览(可滚动查看)