Vision-Language Foundation Models as Effective Robot Imitators

Open paper

Benchmark	Model	Rank	Results
robot-manipulation-on-calvin	Roboflamingo	–	avg. sequence length (D to D): 2.47