I am using Jetson Xavier with TensroRT 5.0.3, I cloned the repo and did all the steps required before I make the environment, my Makefile looks like this :
Makefile.config
CXX=g++
CUDA_VER:=10
#Set to TEGRA for jetson or TESLA for dGPU’s
PLATFORM:=TEGRA
#For Tesla Plugins
OPENCV_INSTALL_DIR:= /path/to/opencv-3.4.x
TENSORRT_INSTALL_DIR:= /path/to/TensorRT-5.x
DEEPSTREAM_INSTALL_DIR:= /path/to/DeepStream_Release_3.0
I tried to deploy trt-yolo-app on jetson Xavier and I followed the instructions listed in the original Github repository. I was able to make the app but when I executed it I got the following error:
nvidia@jetson:~deepstream-plugins$ trt-yolo-app --flagfile=/home/nvidia/Projects/deepstream-plugins/config/deepstream-app_yolo_config.txt
Using previously generated plan file located at sources/lib/models/yolov3-kFLOAT-batch1.engine
Loading TRT Engine...
ERROR: The engine plan file is incompatible with this version of TensorRT, expecting 5.0.3.2got 5.0.2.6, please rebuild.
Loading Complete!
trt-yolo-app: yolo.cpp:100: Yolo::Yolo(uint): Assertion `m_Engine != nullptr' failed.
Aborted (core dumped)
I know that my jetson Xavier has 5.0.3 TensorRT version installed on it, but from where the 5.0.2.6 build is coming, is it the original build provided by the repo and how can I change it?
Another question regarding DeepStream SDK 3.0. I am curious to know if there is an image resize for my input. I am testing the model with 1024 x 636 and 1352 x 900 resolution images and I want to know if there is any transformation before the first CNN. Where in the code I can find this information. I know that YOLO resizes by a step of 32. I am wondering if it is the same for DeepStream.