How to make IResizeLayer work with dynamic batch size

AI & Data Science Deep Learning (Training & Inference) TensorRT
1

TensorRT version 7.1.3.4

    auto layer = net->addResize(*in);
    layer->setOutputDimensions(outShape);

To work with dynamic batch size, the first dimension of outShape is set -1. Then, TensorRT complains “…/builder/cudnnBuilderGraph.cpp (780) - Assertion Error in checkDimsSanity: 0 (dims.d[i] >= 0)”.

So, how can I make IResizeLayer works with dynamic batch size? Or this is not supported?

Thanks

2

Hi @uni19,
Looks like your query has already been answered here.
https://github.com/NVIDIA/TensorRT/issues/783#issuecomment-694296307

Thanks!