Can I use both variable(dynamic) input shape and batching in TensorRT?

Description

Can I use both variable(dynamic) input shape and batching? And what’s the strategy to padding input when batching?

Environment

TensorRT version: lastest