阅读(3533) 赞(10)

如何启用Halide后端以提高效率

2017-10-18 10:08:05 更新

介绍

本教程指导如何使用Halide语言后端在OpenCV深度学习模块中运行模型。Halide是一个开源项目，让我们以易读的格式编写图像处理算法，根据具体设备计划计算，并以相当好的效率进行评估。

Halide项目的官方网站：http : //halide-lang.org/。

最新的效率比较：https：//github.com/opencv/opencv/wiki/DNN-效率

要求

LLVM编译器

注意

LLVM编译可能需要很长时间。

从http://releases.llvm.org/4.0.0/llvm-4.0.0.src.tar.xz下载LLVM源代码。打开包装让llvm_root是源代码的根目录。
创建目录llvm_root / tools / clang
下载Clang与LLVM相同的版本。在我们的例子中，它将来自http://releases.llvm.org/4.0.0/cfe-4.0.0.src.tar.xz。将其打包成llvm_root / tools / clang。请注意，它应该是Clang源代码的根目录。

在Linux上构建LLVM

cd llvm_root
mkdir build && cd build
cmake -DLLVM_ENABLE_TERMINFO = OFF -DLLVM_TARGETS_TO_BUILD =“X86”-DLLVM_ENABLE_ASSERTIONS = ON -DCMAKE_BUILD_TYPE = Release ..
make -j4

在Windows上构建LLVM（开发者命令提示符）

mkdir  path-to-llvm-build  && cd  path-to-llvm-build 
cmake.exe -DLLVM_ENABLE_TERMINFO = OFF -DLLVM_TARGETS_TO_BUILD = X86 -DLLVM_ENABLE_ASSERTIONS = ON -DCMAKE_BUILD_TYPE = Release -DCMAKE_INSTALL_PREFIX =  path-to-llvm-install  -G“Visual Studio 14 Win64” path-to-llvm-src 
MSBuild.exe / m：4 / t：Build / p：Configuration = Release。 INSTALL.vcxproj

注意

path-to-llvm-build并且path-to-llvm-install是不同的目录。

Halide language.

从GitHub存储库，https://github.com/halide/Halide或使用git 下载源代码。根目录将是一个halo_root。

git clone https://github.com/halide/Halide.git

在Linux上构建Halide

cd halo_root
mkdir build && cd build
cmake -DLLVM_DIR = llvm_root / build / lib / cmake / llvm -DCMAKE_BUILD_TYPE = Release -DLLVM_VERSION = 40 -DWITH_TESTS = OFF -DWITH_APPS = OFF -DWITH_TUTORIALS = OFF ..
make -j4

在Windows上构建Halide（开发者命令提示符）

cd halo_root
mkdir build && cd build
cmake.exe -DLLVM_DIR =  path-to-llvm-install  lib  cmake  llvm -DLLVM_VERSION = 40 -DWITH_TESTS = OFF -DWITH_APPS = OFF -DWITH_TUTORIALS = OFF -DCMAKE_BUILD_TYPE = Release -G“Visual Studio 14 Win64“..
MSBuild.exe / m：4 / t：Build / p：Configuration = Release。 ALL_BUILD.vcxproj

用Halide后端构建OpenCV

构建OpenCV时，添加以下配置标志：

ENABLE_CXX11 - 启用C ++ 11标准
WITH_HALIDE - 启用卤化物连接
HALIDE_ROOT_DIR - Halide构建目录的路径

Sample

// Sample of using Halide backend in OpenCV deep learning module.
// Based on caffe_googlenet.cpp.
#include <opencv2/dnn.hpp>
#include <opencv2/imgproc.hpp>
#include <opencv2/highgui.hpp>
using namespace cv;
using namespace cv::dnn;
#include <fstream>
#include <iostream>
#include <cstdlib>
/* Find best class for the blob (i. e. class with maximal probability) */
static void getMaxClass(const Mat &probBlob, int *classId, double *classProb)
{
    Mat probMat = probBlob.reshape(1, 1); //reshape the blob to 1x1000 matrix
    Point classNumber;
    minMaxLoc(probMat, NULL, classProb, NULL, &classNumber);
    *classId = classNumber.x;
}
static std::vector<std::string> readClassNames(const char *filename = "synset_words.txt")
{
    std::vector<std::string> classNames;
    std::ifstream fp(filename);
    if (!fp.is_open())
    {
        std::cerr << "File with classes labels not found: " << filename << std::endl;
        exit(-1);
    }
    std::string name;
    while (!fp.eof())
    {
        std::getline(fp, name);
        if (name.length())
            classNames.push_back( name.substr(name.find(' ')+1) );
    }
    fp.close();
    return classNames;
}
int main(int argc, char **argv)
{
    std::string modelTxt = "train_val.prototxt";
    std::string modelBin = "squeezenet_v1.1.caffemodel";
    std::string imageFile = (argc > 1) ? argv[1] : "space_shuttle.jpg";
    Net net = dnn::readNetFromCaffe(modelTxt, modelBin);
    if (net.empty())
    {
        std::cerr << "Can't load network by using the following files: " << std::endl;
        std::cerr << "prototxt:   " << modelTxt << std::endl;
        std::cerr << "caffemodel: " << modelBin << std::endl;
        std::cerr << "SqueezeNet v1.1 can be downloaded from:" << std::endl;
        std::cerr << "https://github.com/DeepScale/SqueezeNet/tree/master/SqueezeNet_v1.1" << std::endl;
        exit(-1);
    }
    Mat img = imread(imageFile);
    if (img.empty())
    {
        std::cerr << "Can't read image from the file: " << imageFile << std::endl;
        exit(-1);
    }
    if (img.channels() != 3)
    {
        std::cerr << "Image " << imageFile << " isn't 3-channel" << std::endl;
        exit(-1);
    }
    resize(img, img, Size(227, 227));                // SqueezeNet v1.1 predict class by 3x227x227 input image.
    Mat inputBlob = blobFromImage(img, 1.0, Size(), Scalar(), false);  // Convert Mat to 4-dimensional batch.
    net.setInput(inputBlob);                         // Set the network input.
    net.setPreferableBackend(DNN_BACKEND_HALIDE);    // Tell engine to use Halide where it possible.
    Mat prob = net.forward("prob");                  // Compute output.
    int classId;
    double classProb;
    getMaxClass(prob, &classId, &classProb);         // Find the best class.
    std::vector<std::string> classNames = readClassNames();
    std::cout << "Best class: #" << classId << " '" << classNames.at(classId) << "'" << std::endl;
    std::cout << "Probability: " << classProb * 100 << "%" << std::endl;
    return 0;
} //main

说明

从SqueezeNet仓库下载Caffe模型：train_val.prototxt和squeezenet_v1.1.caffemodel。

还需要使用名称为ILSVRC2012类的文件：synset_words.txt。

将这些文件放入此程序示例的工作目录中。

使用.prototxt和.caffemodel文件的路径读取并初始化网络
```
    Net net = dnn::readNetFromCaffe(modelTxt, modelBin);
```

检查网络是否已成功读取

    if (net.empty())
    {
        std::cerr << "Can't load network by using the following files: " << std::endl;
        std::cerr << "prototxt:   " << modelTxt << std::endl;
        std::cerr << "caffemodel: " << modelBin << std::endl;
        std::cerr << "SqueezeNet v1.1 can be downloaded from:" << std::endl;
        std::cerr << "https://github.com/DeepScale/SqueezeNet/tree/master/SqueezeNet_v1.1" << std::endl;
        exit(-1);
    }

读取输入图像并转换为四维Blob，SqueezeNet v1.1可以接受

    Mat img = imread(imageFile);
    if (img.empty())
    {
        std::cerr << "Can't read image from the file: " << imageFile << std::endl;
        exit(-1);
    }
    if (img.channels() != 3)
    {
        std::cerr << "Image " << imageFile << " isn't 3-channel" << std::endl;
        exit(-1);
    }
    resize(img, img, Size(227, 227));                // SqueezeNet v1.1 predict class by 3x227x227 input image.
    Mat inputBlob = blobFromImage(img, 1.0, Size(), Scalar(), false);  // Convert Mat to 4-dimensional batch.

将blob传递到网络

    net.setInput(inputBlob);                         // Set the network input.

为其实现的层启用Halide后端

    net.setPreferableBackend(DNN_BACKEND_HALIDE);    // Tell engine to use Halide where it possible.

Make forward pass
- 记住，初始化之后的第一个向前传递需要相当多的时间来实现下一个。这是因为第一次调用时Halide管道的运行时编译。
```
    Mat prob = net.forward("prob");                  // Compute output.
```
确定最好的班级

    int classId;
    double classProb;
    getMaxClass(prob, &classId, &classProb);         // Find the best class.

打印结果

    std::vector<std::string> classNames = readClassNames();
    std::cout << "Best class: #" << classId << " '" << classNames.at(classId) << "'" << std::endl;
    std::cout << "Probability: " << classProb * 100 << "%" << std::endl;

对于我们的图像我们得到：

Best class: #812 'space shuttle'

Probability: 97.9812%

← 加载Caffe框架模型

如何为Halide后端安排网络 →

如何启用Halide后端以提高效率

介绍

要求

LLVM编译器

Halide language.

用Halide后端构建OpenCV

Sample

说明

推荐文章

推荐教程

最新教程