『heqingchun-Ubuntu系统+x86架构+配置编译安装使用yolov5-6.0+带有TensorRT硬件加速+封装动态库+C++部署+Qt』

本文主要是介绍『heqingchun-Ubuntu系统+x86架构+配置编译安装使用yolov5-6.0+带有TensorRT硬件加速+封装动态库+C++部署+Qt』,希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!

Ubuntu系统+x86架构+配置编译安装使用yolov5-6.0+带有TensorRT硬件加速+封装动态库+C++部署+Qt

学习本篇文章后你会:部署yolov5-6.0版本、模型转换(.pt->.wts->.engine)、将yolov5-6.0封装动态库并得到深度学习推理结果,最后在qt或其他项目中调用使用。

一、准备文件

1.yolov5-6.0.zip

官网下载
网址:

https://github.com/ultralytics/yolov5/tree/v6.0

操作:
点击"Code"下的"Download ZIP"
下载得到yolov5-6.0.zip压缩文件

2.tensorrtx-yolov5-v6.0.zip

官网下载
网址:

https://github.com/wang-xinyu/tensorrtx/tree/yolov5-v6.0

操作:
点击"Code"下的"Download ZIP"
下载得到tensorrtx-yolov5-v6.0.zip压缩文件

3.yolov5s.pt

官网下载
网址:

https://github.com/ultralytics/yolov5/tree/v6.0

向页面查找"YOLOv5s"

YOLOv5n 	640 	28.4 	46.0 	45 	6.3 	0.6 	1.9 	4.5
YOLOv5s 	640 	37.2 	56.0 	98 	6.4 	0.9 	7.2 	16.5
YOLOv5m 	640 	45.2 	63.9 	224 	8.2 	1.7 	21.2 	49.0
YOLOv5l 	640 	48.8 	67.2 	430 	10.1 	2.7 	46.5 	109.1
YOLOv5x 	640 	50.7 	68.9 	766 	12.1 	4.8 	86.7 	205.7				
YOLOv5n6 	1280 	34.0 	50.7 	153 	8.1 	2.1 	3.2 	4.6
YOLOv5s6 	1280 	44.5 	63.0 	385 	8.2 	3.6 	16.8 	12.6
YOLOv5m6 	1280 	51.0 	69.0 	887 	11.1 	6.8 	35.7 	50.0
YOLOv5l6 	1280 	53.6 	71.6 	1784 	15.8 	10.5 	76.8 	111.4

操作:
点击"YOLOv5s"即可下载yolov5s.pt文件

4.将文件按以下顺序存放

新建"TensorRT"目录

mkdir TensorRT

"yolov5-6.0.zip"放入TensorRT目录
"tensorrtx-yolov5-v6.0.zip"放入TensorRT目录
"yolov5s.pt"放入TensorRT目录
存放好后将压缩文件解压即可

二、更新、安装基础依赖

sudo apt update && \
sudo apt upgrade -y && \
sudo apt install -y build-essential cmake pip

三、安装依赖

1.nvidia驱动、cuda、cudnn、tensorRT

参考我的博客: 『heqingchun-ubuntu系统下安装nvidia显卡驱动3种方法』
参考我的博客: 『heqingchun-ubuntu系统下安装cuda与cudnn』
参考我的博客: 『heqingchun-ubuntu使用TensorRT配置』

2.pytorch

先设置pip加速,要不然非常慢: Ubuntu系统+设置pip加速
安装pytorch前设置环境变量
打开

sudo gedit /etc/profile

写入

export PATH=/home/heqingchun/.local/bin:$PATH

更新

source /etc/profile

参考我的博客: ubuntu开发环境配置(cuda、cudnn、ffmpeg、opencv、darknet-master、TensorRT、python、pytorch、MySql、qt(armv8交叉编译))
第"九"项

CUDA 11.8
pip3 install torch==2.1.0 torchvision==0.16.0 torchaudio==2.1.0 --index-url https://download.pytorch.org/whl/cu118

3.requirements.txt

解压下载的“yolov5-6.0.zip”压缩文件,进入目录

unzip yolov5-6.0.zip 
cd yolov5-6.0

执行:

pip3 install -r requirements.txt

如果遇到问题:ERROR: pandas 2.0.3 has requirement python-dateutil>=2.8.2, but you’ll have python-dateutil 2.7.3 which is incompatible.则执行:

pip3 install --upgrade python-dateutil

四、开始模型转换

1.yolov5.pt转换为yolov5.wts

将yolov5.pt文件与tensorrtx-yolov5-v6.0/yolov5目录下的gen_wts.py文件放置到yolov5-6.0目录下

cd TensorRT
cp yolov5s.pt tensorrtx-yolov5-v6.0/yolov5/gen_wts.py yolov5-6.0

转换

python3 yolov5-6.0/gen_wts.py --w yolov5-6.0/yolov5s.pt --o yolov5s.wts

在TensorRT目录生成yolov5s.wts文件

2.yolov5s.wts转换为yolov5s.engine

修改cmake文件

cd TensorRT
gedit tensorrtx-yolov5-v6.0/yolov5/CMakeLists.txt

修改tensorrt头文件与库文件目录为当前正确目录
原:

# tensorrt
include_directories(/usr/include/x86_64-linux-gnu/)
link_directories(/usr/lib/x86_64-linux-gnu/)

新:

# tensorrt
include_directories(/home/heqingchun/soft/TensorRT/TensorRT-8.4.3.1/include)
link_directories(/home/heqingchun/soft/TensorRT/TensorRT-8.4.3.1/lib)

新建build目录

mkdir -p tensorrtx-yolov5-v6.0/yolov5/build

进入build目录

cd tensorrtx-yolov5-v6.0/yolov5/build

编译

cmake ..
make -j $(nproc)

在build目录会生成可执行文件,可用于转换模型与推理检测。注:当前使用模型文件为官方模型,故检测类型数量为80,如需更改,请在cmake前修改yololayer.h中的num_classes 的数量。
转换

./yolov5 -s ../../../yolov5s.wts yolov5s.engine s

等待一会就会发现已经成功转换出yolov5s.engine文件了。

五、推理测试

cd TensorRT/tensorrtx-yolov5-v6.0/yolov5/build/
./yolov5 -d yolov5s.engine ../samples

控制台打印信息:

[11/27/2023-09:37:24] [W] [TRT] CUDA lazy loading is not enabled. Enabling it can significantly reduce device memory usage. See `CUDA_MODULE_LOADING` in https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#env-vars
inference time: 3ms
inference time: 3ms

在samples目录中存放有2张图片
可见推理成功,并生成两张图片"_bus.jpg"与"_zidane.jpg"已画框,时间为3ms。

六、封装后在C++调用

在以上操作均没问题的情况下,目前已经能使用生成的yolov5进行深度学习推理检测了,不过只能在控制台进行命令行操作,接下来进行动态库的封装以及将库嵌入到qt项目中进行测试,以方便后续使用。

1.将"build"目录拷贝一份,命名为"YOLOv5v60"作为区分

cd TensorRT/tensorrtx-yolov5-v6.0
cp -fr yolov5 YOLOv5v60

2.新建源文件"YOLOv5v60.cpp"

cd YOLOv5v60
touch YOLOv5v60.cpp
gedit YOLOv5v60.cpp

写入以下内容并保存

#include <iostream>
#include <fstream>
#include <sstream>
#include <string.h>
#include <vector>
#include <chrono>
#include "cuda_runtime_api.h"
#include "logging.h"
#include "common.hpp"
#include "calibrator.h"
#include "cuda_utils.h"
#include <opencv2/opencv.hpp>
#include <opencv2/dnn.hpp>
#define USE_FP16  // set USE_INT8 or USE_FP16 or USE_FP32
#define DEVICE 0  // GPU id
#define BATCH_SIZE 1
using namespace std;
using namespace cv;
static const int INPUT_H = Yolo::INPUT_H;
static const int INPUT_W = Yolo::INPUT_W;
static const int CLASS_NUM = Yolo::CLASS_NUM;
static const int OUTPUT_SIZE = Yolo::MAX_OUTPUT_BBOX_COUNT * sizeof(Yolo::Detection) / sizeof(float) + 1; 
const char* INPUT_BLOB_NAME = "data";
const char* OUTPUT_BLOB_NAME = "prob";
static Logger gLogger;
typedef struct {int classid; // 标签名idcv::Rect rbox; // 框float conf; // 置信度
} YOLOv5v60Result;
typedef struct {float *data;float *prob;IExecutionContext *exe_context;void* buffers[2];cudaStream_t cuda_stream;int inputIndex;int outputIndex;
} YOLOv5v60TRTContext;static inline cv::Mat PreprocessImage(cv::Mat& img, int input_w, int input_h) {int w, h, x, y;float r_w = input_w / (img.cols*1.0);float r_h = input_h / (img.rows*1.0);if (r_h > r_w) {w = input_w;h = r_w * img.rows;x = 0;y = (input_h - h) / 2;} else {w = r_h * img.cols;h = input_h;x = (input_w - w) / 2;y = 0;}cv::Mat re(h, w, CV_8UC3);cv::cvtColor(img,img,cv::COLOR_RGBA2RGB);cv::resize(img, re, re.size(), 0, 0, cv::INTER_LINEAR);cv::Mat out(input_h, input_w, CV_8UC3, cv::Scalar(128, 128, 128));re.copyTo(out(cv::Rect(x, y, re.cols, re.rows)));return out;
}extern "C" YOLOv5v60TRTContext* InitEngine(const char* enginePath);
extern "C" void YOLOv5v60TRTDetect(YOLOv5v60TRTContext *trt_ctx,cv::Mat img, std::vector<YOLOv5v60Result>& vYoloresult, float conf_thresh, float nms_thresh);
extern "C" void DeleteYolo(YOLOv5v60TRTContext *trt_ctx);
static void doInference(IExecutionContext& context, cudaStream_t& stream, void **buffers, float* input, float* output, int batchSize) 
{// DMA input batch data to device, infer on the batch asynchronously, and DMA output back to hostCUDA_CHECK(cudaMemcpyAsync(buffers[0], input, batchSize * 3 * INPUT_H * INPUT_W * sizeof(float), cudaMemcpyHostToDevice, stream));context.enqueue(batchSize, buffers, stream, nullptr);CUDA_CHECK(cudaMemcpyAsync(output, buffers[1], batchSize * OUTPUT_SIZE * sizeof(float), cudaMemcpyDeviceToHost, stream));cudaStreamSynchronize(stream);
}
extern "C" YOLOv5v60TRTContext* InitEngine(const char* enginePath)
{size_t size;char *trtModelStream = NULL;YOLOv5v60TRTContext * trt_ctx = NULL;trt_ctx = new YOLOv5v60TRTContext();trt_ctx->data = new float[BATCH_SIZE * 3 * INPUT_H * INPUT_W];trt_ctx->prob = new float[BATCH_SIZE * OUTPUT_SIZE];std::ifstream file(enginePath, std::ios::binary);if (file.good()){file.seekg(0, file.end);size = file.tellg();file.seekg(0, file.beg);trtModelStream = new char[size];assert(trtModelStream);file.read(trtModelStream, size);file.close();}IRuntime* runtime = createInferRuntime(gLogger);assert(runtime != nullptr);ICudaEngine *engine = runtime->deserializeCudaEngine(trtModelStream, size);assert(engine != nullptr);trt_ctx->exe_context = engine->createExecutionContext();assert(trt_ctx->exe_context != nullptr);delete[] trtModelStream;assert(engine->getNbBindings() == 2);// In order to bind the buffers, we need to know the names of the input and output tensors.// Note that indices are guaranteed to be less than IEngine::getNbBindings()const int inputIndex = engine->getBindingIndex(INPUT_BLOB_NAME);const int outputIndex = engine->getBindingIndex(OUTPUT_BLOB_NAME);assert(inputIndex == 0);assert(outputIndex == 1);CUDA_CHECK(cudaMalloc(&trt_ctx->buffers[inputIndex], BATCH_SIZE * 3 * INPUT_H * INPUT_W * sizeof(float)));CUDA_CHECK(cudaMalloc(&trt_ctx->buffers[outputIndex], BATCH_SIZE * OUTPUT_SIZE * sizeof(float)));CUDA_CHECK(cudaStreamCreate(&trt_ctx->cuda_stream));printf("YOLOv5v60 InitEngine successed\n");return (YOLOv5v60TRTContext *)trt_ctx;
}
extern "C" void YOLOv5v60TRTDetect(YOLOv5v60TRTContext *trt_ctx,cv::Mat img, std::vector<YOLOv5v60Result>& vYoloresult, float conf_thresh, float nms_thresh)
{printf("YOLOv5v60TRTDetect start\n");cv::Mat pr_img = PreprocessImage(img, INPUT_W, INPUT_H);// letterbox BGR to RGBint i = 0;for (int row = 0; row < INPUT_H; ++row){uchar* uc_pixel = pr_img.data + row * pr_img.step;for (int col = 0; col < INPUT_W; ++col){trt_ctx->data[i] = (float)uc_pixel[2] / 255.0;trt_ctx->data[i + INPUT_H * INPUT_W] = (float)uc_pixel[1] / 255.0;trt_ctx->data[i + 2 * INPUT_H * INPUT_W] = (float)uc_pixel[0] / 255.0;uc_pixel += 3;++i;}}// Run inferenceauto start = std::chrono::system_clock::now();doInference(*trt_ctx->exe_context, trt_ctx->cuda_stream, trt_ctx->buffers, trt_ctx->data, trt_ctx->prob, BATCH_SIZE);auto end = std::chrono::system_clock::now();std::cout << "检测耗时:" << std::chrono::duration_cast<std::chrono::milliseconds>(end - start).count() << "ms" << std::endl;std::vector<Yolo::Detection> batch_res;nms(batch_res, &trt_ctx->prob[0], conf_thresh, nms_thresh);for (size_t j = 0; j < batch_res.size(); j++){cv::Rect r = get_rect(img, batch_res[j].bbox);YOLOv5v60Result yoloresult;yoloresult.rbox = r;yoloresult.classid = (int)batch_res[j].class_id;yoloresult.conf = ((float)(int)((batch_res[j].conf + 0.005) * 100)) / 100;vYoloresult.push_back(yoloresult);}
}
extern "C" void DeleteYolo(YOLOv5v60TRTContext *trt_ctx)
{cudaStreamDestroy(trt_ctx->cuda_stream);CUDA_CHECK(cudaFree(trt_ctx->buffers[0]));CUDA_CHECK(cudaFree(trt_ctx->buffers[1]));trt_ctx->exe_context->destroy();delete trt_ctx->data;delete trt_ctx->prob;delete trt_ctx;
}

3.修改"CMakeLists.txt"文件

修改位置我已做好标记(#heqingchun)。

原文件:

cmake_minimum_required(VERSION 2.6)project(yolov5)add_definitions(-std=c++11)
add_definitions(-DAPI_EXPORTS)
option(CUDA_USE_STATIC_CUDA_RUNTIME OFF)
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_BUILD_TYPE Debug)find_package(CUDA REQUIRED)if(WIN32)
enable_language(CUDA)
endif(WIN32)include_directories(${PROJECT_SOURCE_DIR}/include)
# include and link dirs of cuda and tensorrt, you need adapt them if yours are different
# cuda
include_directories(/usr/local/cuda/include)
link_directories(/usr/local/cuda/lib64)
# tensorrt
include_directories(/home/heqingchun/soft/TensorRT/TensorRT-8.5.3.1/include)
link_directories(/home/heqingchun/soft/TensorRT/TensorRT-8.5.3.1/lib)set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wall -Ofast -g -Wfatal-errors -D_MWAITXINTRIN_H_INCLUDED")
cuda_add_library(myplugins SHARED yololayer.cu)
target_link_libraries(myplugins nvinfer cudart)find_package(OpenCV)
include_directories(${OpenCV_INCLUDE_DIRS})cuda_add_executable(yolov5 calibrator.cpp yolov5.cpp preprocess.cu)target_link_libraries(yolov5 nvinfer)
target_link_libraries(yolov5 cudart)
target_link_libraries(yolov5 myplugins)
target_link_libraries(yolov5 ${OpenCV_LIBS})if(UNIX)
add_definitions(-O2 -pthread)
endif(UNIX)

修改后文件:

cmake_minimum_required(VERSION 2.6)#heqingchun project(yolov5)
#heqingchun
project(YOLOv5v60)add_definitions(-std=c++11)
add_definitions(-DAPI_EXPORTS)
option(CUDA_USE_STATIC_CUDA_RUNTIME OFF)
set(CMAKE_CXX_STANDARD 11)
set(CMAKE_BUILD_TYPE Debug)find_package(CUDA REQUIRED)if(WIN32)
enable_language(CUDA)
endif(WIN32)include_directories(${PROJECT_SOURCE_DIR}/include)
# include and link dirs of cuda and tensorrt, you need adapt them if yours are different
# cuda
include_directories(/usr/local/cuda/include)
link_directories(/usr/local/cuda/lib64)
# tensorrt
include_directories(/home/heqingchun/soft/TensorRT/TensorRT-8.5.3.1/include)
link_directories(/home/heqingchun/soft/TensorRT/TensorRT-8.5.3.1/lib)set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -std=c++11 -Wall -Ofast -g -Wfatal-errors -D_MWAITXINTRIN_H_INCLUDED")
#heqingchun cuda_add_library(myplugins SHARED yololayer.cu)
#heqingchun target_link_libraries(myplugins nvinfer cudart)
#heqingchun
cuda_add_library(YOLOv5v60 SHARED yololayer.cu YOLOv5v60.cpp)find_package(OpenCV)
include_directories(${OpenCV_INCLUDE_DIRS})#heqingchun cuda_add_executable(yolov5 calibrator.cpp yolov5.cpp preprocess.cu)#heqingchun target_link_libraries(yolov5 nvinfer)
#heqingchun target_link_libraries(yolov5 cudart)
#heqingchun target_link_libraries(yolov5 myplugins)
#heqingchun target_link_libraries(yolov5 ${OpenCV_LIBS})
#heqingchun
target_link_libraries(YOLOv5v60 nvinfer)
target_link_libraries(YOLOv5v60 cudart)
target_link_libraries(YOLOv5v60 ${OpenCV_LIBS})if(UNIX)
add_definitions(-O2 -pthread)
endif(UNIX)

修改后保存

4.编译动态库

进入build目录,将文件全部删除

cd build
rm -fr *

编译

cmake ..
make -j $(nproc)

结束会在build目录生成"libYOLOv5v60.so"动态库文件

5.使用动态库

打开qt,新建mainwindow桌面程序,ui界面增加一个按钮,并准备好槽函数,qt项目文件如下:

(1)YOLOv5v60.pro
QT       += core guigreaterThan(QT_MAJOR_VERSION, 4): QT += widgetsCONFIG += c++11# The following define makes your compiler emit warnings if you use
# any Qt feature that has been marked deprecated (the exact warnings
# depend on your compiler). Please consult the documentation of the
# deprecated API in order to know how to port your code away from it.
DEFINES += QT_DEPRECATED_WARNINGS# You can also make your code fail to compile if it uses deprecated APIs.
# In order to do so, uncomment the following line.
# You can also select to disable deprecated APIs only up to a certain version of Qt.
#DEFINES += QT_DISABLE_DEPRECATED_BEFORE=0x060000    # disables all the APIs deprecated before Qt 6.0.0SOURCES += \main.cpp \mainwindow.cppHEADERS += \mainwindow.hFORMS += \mainwindow.ui# Default rules for deployment.
qnx: target.path = /tmp/$${TARGET}/bin
else: unix:!android: target.path = /opt/$${TARGET}/bin
!isEmpty(target.path): INSTALLS += targetINCLUDEPATH += /usr/local/include/opencv4
unix|win32: LIBS += -L/usr/local/lib/ -lopencv_worldINCLUDEPATH += /usr/local/cuda/include
unix|win32: LIBS += -L/usr/local/cuda/lib64/ -lcudartINCLUDEPATH += /home/heqingchun/soft/TensorRT/TensorRT-8.5.3.1/include
unix|win32: LIBS += -L$$PWD/../../../soft/TensorRT/TensorRT-8.5.3.1/lib/ -lnvinferINCLUDEPATH += /home/heqingchun/soft/TensorRT/tensorrtx-yolov5-v6.0/YOLOv5v60
unix|win32: LIBS += -L$$PWD/../../../soft/TensorRT/tensorrtx-yolov5-v6.0/YOLOv5v60/build/ -lYOLOv5v60
(2)main.cpp(没有编辑)
#include "mainwindow.h"#include <QApplication>int main(int argc, char *argv[])
{QApplication a(argc, argv);MainWindow w;w.show();return a.exec();
}
(3)mainwindow.h
#ifndef MAINWINDOW_H
#define MAINWINDOW_H#include <QMainWindow>
#include <QDebug>QT_BEGIN_NAMESPACE
namespace Ui { class MainWindow; }
QT_END_NAMESPACEclass MainWindow : public QMainWindow
{Q_OBJECTpublic:MainWindow(QWidget *parent = nullptr);~MainWindow();private slots:void on_pushButton_clicked();private:Ui::MainWindow *ui;
};
#endif // MAINWINDOW_H
(4)mainwindow.cpp
#include "mainwindow.h"
#include "ui_mainwindow.h"#include "YOLOv5v60.cpp"
extern "C" YOLOv5v60TRTContext* InitEngine(const char* enginePath);
extern "C" void YOLOv5v60TRTDetect(YOLOv5v60TRTContext *trt_ctx,cv::Mat img, std::vector<YOLOv5v60Result>& vYoloresult, float conf_thresh, float nms_thresh);
extern "C" void DeleteYolo(YOLOv5v60TRTContext *trt_ctx);YOLOv5v60TRTContext * trt_ctx = NULL;
vector<YOLOv5v60Result> vYoloresult;
cv::Mat mat;
QStringList nameList = {"person", "bicycle", "car", "motorcycle", "airplane", "bus", "train", "truck", "boat", "traffic light",
"fire hydrant", "stop sign", "parking meter", "bench", "bird", "cat", "dog", "horse", "sheep", "cow",
"elephant", "bear", "zebra", "giraffe", "backpack", "umbrella", "handbag", "tie", "suitcase", "frisbee",
"skis", "snowboard", "sports ball", "kite", "baseball bat", "baseball glove", "skateboard", "surfboard",
"tennis racket", "bottle", "wine glass", "cup", "fork", "knife", "spoon", "bowl", "banana", "apple",
"sandwich", "orange", "broccoli", "carrot", "hot dog", "pizza", "donut", "cake", "chair", "couch",
"potted plant", "bed", "dining table", "toilet", "tv", "laptop", "mouse", "remote", "keyboard", "cell phone",
"microwave", "oven", "toaster", "sink", "refrigerator", "book", "clock", "vase", "scissors", "teddy bear",
"hair drier", "toothbrush"};MainWindow::MainWindow(QWidget *parent): QMainWindow(parent), ui(new Ui::MainWindow)
{ui->setupUi(this);std::string modePath = "/home/heqingchun/soft/TensorRT/tensorrtx-yolov5-v6.0/yolov5/build/yolov5s.engine";trt_ctx = InitEngine(modePath.data());
}MainWindow::~MainWindow()
{delete ui;
}void MainWindow::on_pushButton_clicked()
{mat = cv::imread("/home/heqingchun/soft/TensorRT/tensorrtx-yolov5-v6.0/yolov5/samples/bus.jpg");YOLOv5v60TRTDetect(trt_ctx,mat,vYoloresult,0.6,0.45);for (unsigned long i = 0; i < vYoloresult.size(); i++) {YOLOv5v60Result result = vYoloresult[i];QString classId =nameList[result.classid];qDebug()<<"current classes:"<<classId<<"conf:"<<result.conf<<result.rbox.x<<result.rbox.y<<result.rbox.width<<result.rbox.height;// 画框cv::rectangle(mat,cv::Point(result.rbox.x,result.rbox.y),cv::Point(result.rbox.x + result.rbox.width,result.rbox.y + result.rbox.height),cv::Scalar(0,0,255));// 写字cv::putText(mat,classId.toStdString() + ":" + QString::number(result.conf).toStdString(),cv::Point(result.rbox.x,result.rbox.y),5,1,cv::Scalar(0,0,255));}cv::imshow("结果",mat);mat.release();vYoloresult.clear();
}

Ubuntu系统+x86架构+配置编译安装使用yolov5-6.0+带有TensorRT硬件加速+封装动态库+C++部署+Qt-完毕

这篇关于『heqingchun-Ubuntu系统+x86架构+配置编译安装使用yolov5-6.0+带有TensorRT硬件加速+封装动态库+C++部署+Qt』的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!



http://www.chinasem.cn/article/455318

相关文章

C++使用栈实现括号匹配的代码详解

《C++使用栈实现括号匹配的代码详解》在编程中,括号匹配是一个常见问题,尤其是在处理数学表达式、编译器解析等任务时,栈是一种非常适合处理此类问题的数据结构,能够精确地管理括号的匹配问题,本文将通过C+... 目录引言问题描述代码讲解代码解析栈的状态表示测试总结引言在编程中,括号匹配是一个常见问题,尤其是在

Java中String字符串使用避坑指南

《Java中String字符串使用避坑指南》Java中的String字符串是我们日常编程中用得最多的类之一,看似简单的String使用,却隐藏着不少“坑”,如果不注意,可能会导致性能问题、意外的错误容... 目录8个避坑点如下:1. 字符串的不可变性:每次修改都创建新对象2. 使用 == 比较字符串,陷阱满

Python使用国内镜像加速pip安装的方法讲解

《Python使用国内镜像加速pip安装的方法讲解》在Python开发中,pip是一个非常重要的工具,用于安装和管理Python的第三方库,然而,在国内使用pip安装依赖时,往往会因为网络问题而导致速... 目录一、pip 工具简介1. 什么是 pip?2. 什么是 -i 参数?二、国内镜像源的选择三、如何

使用C++实现链表元素的反转

《使用C++实现链表元素的反转》反转链表是链表操作中一个经典的问题,也是面试中常见的考题,本文将从思路到实现一步步地讲解如何实现链表的反转,帮助初学者理解这一操作,我们将使用C++代码演示具体实现,同... 目录问题定义思路分析代码实现带头节点的链表代码讲解其他实现方式时间和空间复杂度分析总结问题定义给定

IDEA编译报错“java: 常量字符串过长”的原因及解决方法

《IDEA编译报错“java:常量字符串过长”的原因及解决方法》今天在开发过程中,由于尝试将一个文件的Base64字符串设置为常量,结果导致IDEA编译的时候出现了如下报错java:常量字符串过长,... 目录一、问题描述二、问题原因2.1 理论角度2.2 源码角度三、解决方案解决方案①:StringBui

Linux使用nload监控网络流量的方法

《Linux使用nload监控网络流量的方法》Linux中的nload命令是一个用于实时监控网络流量的工具,它提供了传入和传出流量的可视化表示,帮助用户一目了然地了解网络活动,本文给大家介绍了Linu... 目录简介安装示例用法基础用法指定网络接口限制显示特定流量类型指定刷新率设置流量速率的显示单位监控多个

ElasticSearch+Kibana通过Docker部署到Linux服务器中操作方法

《ElasticSearch+Kibana通过Docker部署到Linux服务器中操作方法》本文介绍了Elasticsearch的基本概念,包括文档和字段、索引和映射,还详细描述了如何通过Docker... 目录1、ElasticSearch概念2、ElasticSearch、Kibana和IK分词器部署

部署Vue项目到服务器后404错误的原因及解决方案

《部署Vue项目到服务器后404错误的原因及解决方案》文章介绍了Vue项目部署步骤以及404错误的解决方案,部署步骤包括构建项目、上传文件、配置Web服务器、重启Nginx和访问域名,404错误通常是... 目录一、vue项目部署步骤二、404错误原因及解决方案错误场景原因分析解决方案一、Vue项目部署步骤

JavaScript中的reduce方法执行过程、使用场景及进阶用法

《JavaScript中的reduce方法执行过程、使用场景及进阶用法》:本文主要介绍JavaScript中的reduce方法执行过程、使用场景及进阶用法的相关资料,reduce是JavaScri... 目录1. 什么是reduce2. reduce语法2.1 语法2.2 参数说明3. reduce执行过程

如何使用Java实现请求deepseek

《如何使用Java实现请求deepseek》这篇文章主要为大家详细介绍了如何使用Java实现请求deepseek功能,文中的示例代码讲解详细,感兴趣的小伙伴可以跟随小编一起学习一下... 目录1.deepseek的api创建2.Java实现请求deepseek2.1 pom文件2.2 json转化文件2.2