Adding Tensorrt Plugin to Deserialized Engine

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

I have a YOLO network in Caffe format, building the network as engine is succesful and I can infer with it. However I also want to migrate the preprocessing part (resizing and cropping) to TRT too. Guess i need to use cropAndResize Plugin but couldnt find a way to “attach” it to my pipeline.

Thanks in advance.

Environment

TensorRT Version: 7.0.0.11
GPU Type: RTX 2070 Mobile
Nvidia Driver Version: 460
CUDA Version: 10.0
CUDNN Version: 7.6
Operating System + Version: Ubuntu 18.04

#2

Hi @orcdnz,

Hope following links will help you.

https://github.com/NVIDIA/TensorRT/blob/master/samples/opensource/samplePlugin/README.md

Thank you.

#3

Hi, Request you to check the below reference links for custom plugin implementation.
https://github.com/NVIDIA/TensorRT/tree/master/samples/opensource/sampleOnnxMnistCoordConvAC

Thanks!

#4

Thanks but i was actually looking for a way to use plugins that are already existing in TRT. Preferably without modifying the CAFFE model itself. Any further recommendations?

#5

Hi @orcdnz,

Please refer the following doc.

Thank you.