Robotics

Fast head, slow backbone

mediummcq

Practice question

Why do several VLA systems run a large vision-language backbone at a few Hz while a small action head/decoder runs much faster?