Found 1 document(s)
Page 1 of 1
-
Many manipulation tasks require a robot to remember what it has done and what it has seen. Frontier vision-language models can plan such tasks without task-specific training, but how such a planner should remember an epi...