This paper studies imitation learning in the goal-conditioned settings and the proposed method leverages previous ones in a sensible way. The studied problem is of interest to the community, and the reported results are encouraging.