Tensorflow enable native TensorRT support from version 1.7.0, I verified on GTX1080Ti, inception-resnet v2 fp32 inference get 2x speedup without any accuracy loss. I would like to know whether MXNET have plan to support TensorRT?
@hariag @ThomasDelteil Checkout incubator-mxnet/tests/python/tensorrt/
for examples of using mxnet-tensorrt backend directly.