算法开发：将算法模型编译为目标文件.o或.so，保证模型安全

喜欢天晴

已于 2023-08-30 11:46:09 修改

阅读量197

点赞数

分类专栏：机器学习算法部署落地专栏文章标签：算法人工智能深度学习 c++ YOLO

于 2023-08-30 11:34:47 首次发布

本文链接：https://blog.csdn.net/suanfaxiazai/article/details/132578494

版权

机器学习算法部署落地专栏专栏收录该内容

19 篇文章 0 订阅 ¥299.90 ¥99.00

订阅专栏

超级会员免费看

一、环境：jetson orin nx ; yolovx; tensorrt

二、将engine文件转化为.o文件

aarch64-linux-gnu-objcopy -I binary -O elf64-littleaarch64 -B aarch64 model_trt.engine model_trt.o

aarch64-linux-gnu-nm model_trt.o

_binary_model_trt_engine_end
_binary_model_trt_engine_size
_binary_model_trt_engine_start

三、打包为.so文件

cmake_minimum_required(VERSION 2.6)

project(yolox)

add_definitions(-std=c++11)

option(CUDA_USE_STATIC_CUDA_RUNTIME OFF)
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_BUILD_TYPE Debug)

find_package(CUDA REQUIRED)

include_directories(${PROJECT_SOURCE_DIR}/include)
# include and link dirs of cuda and tensorrt, you need adapt them if yours are different
# cuda
# include_directories(/data/cuda/cuda-10.2/cuda/include)
# link_directories(/data/cuda/cuda-10.2/cuda/lib64)
include_directories(/usr/local/cuda-11.4/include)
link_directories(/usr/local/cuda-11.4/lib64)

# cudnn
#include_d