Implementation of batch size inference

[ou525]

Thanks for sharing open source。 While there are differences in batch size inference performance mentioned in the readme, I noticed in the inference code only the provision of multiple batch variables for input images. I haven't seen the implementation of multi-batch for subsequent inference and decoding results. Is there any plan for future support?

[lynettez]

Sorry, there is no plan for that since this sample aims to show how to use cuDLA API and DLA QAT workflow. Thanks for the feedback, anyway~