《看黄色aa大片》

在Altera SoC DE1板卡上跑完整的卷积神经网络

这次为大家详细展示一个利用卷积神经网络实现图片自动分类的例程。

神经网络的优点：自动从数据中学习经验知识，无需复杂的模(🐙)型和算法。

缺点：有监督学习(😈)，需要大量的带标签数据；参数量太少时容易过拟合，泛化能力差，参数量太大时训练收敛很慢（有可能需要几个月到几年）。

为了克服(👕)上述缺点，人们发掘了各种计算资源，包括多核(🎭)CPU、GPU、DSP、ASIC、(🖥)FPGA，甚至使用(🐜)模拟电(📎)路(💇)。

使用CPU实现卷积神经网络比较方便调试，但性能太差，一般人们都选用更快的GPU实现。目前开源的框架大多都支(📆)持GPU，如伯克利大学Caffe和Google Convnet。

微软在2015年2月宣布使用Stratix V完成了CNN加速器，处理 CIFAR10 图片速度可达每秒2300多张。

这里我们也(🕊)使用CIFAR10图片数据，在Cyclone V板子上跑一个卷积神经网络CNN demo。由于板子上计算资源太少（DSP Slice只有80多个），实现完整的网络不太现实，只能在FPGA上实现基本计算单元，然后由HPS统一调度。性能预期不会太(👴)高(📻)，后面给出。

CIFAR10图片都是什么呢？先来张图！

有兴趣的朋友可以到官网下载（(🔘)CIFAR10官网）。上面提(🗑)到过，CNN是有监督学习系统，需要大量带label的数据，CIFAR10就是这样一个开放的数据库，提供了60000张不同类别的图片，分为10个类（如上图左侧所示），每个类别有600张图。这个数据(🙇)集不算特别大，适合(💏)在嵌入式平台上实现。而更大的数据集有ImageNet-1000（ImageNet官网），拥有120多万张高清无码大图，我下载到硬盘，占用了近200GB空间（只能忍痛将其他rmvb和avi删掉了）！

有(🍒)朋友会问，不用这些数据行不行，我们的智能手机里面照片能不能用于CNN做训练？

答案(🔄)是可以的，只(🐄)是你的数据集很不“均匀”，采样不够“完备”，训练出的模型是真实模型的“有偏估(🔌)计”，而上述两个数据集经过了种种(🏖)考验，已经是学术界(🙎)公认的优质数据集，一年一度的ILSVRC比赛就采用了这些数据集。

说完数据，再说模型。先来(✂)看一张经(🌑)典的CNN结构(🍣)：

这是世界上第一个将CNN实用化的例子，实现(🤘)了手写体字母自动识别。在这个CNN模型中，可以看到输入是一张32 x 32的二维图像，经过卷积层(Convolution)、下采样层（Subsampling，也称Pooling)、全连接(🥧)层（Full Connection，也称Inner Product)后，得到一组概率密度，我们选其中概率最大(🥔)的元素作为该模型对输入图像的分类结果。所以实现CNN时，只需要实现三种基本算法：卷积、下采样、矩阵乘。除此(🔍)之(🐢)外，每层输出都(🎫)可选择是否经过非线性变换，常用的非线性变换有ReLU和Sigmoid，前者计算(👍)较为简单，使用较为广泛。

Caffe框架中提供了专门为cifar10数据定制的模型，是(🌨)用proto格式写的，我们的(🥧)demo也基于这个模型。内容如下：

name: "CIFAR10_quick_test"input: "data"input_dim: 1input_dim: 3input_dim: 32input_dim: 32layers {name: "conv1"type: CONVOLUTIONbottom: "data"top: "conv1"blobs_lr: 1blobs_lr: 2convolution_param {num_output: 32pad: 2kernel_size: 5stride: 1}}layers {name: "pool1"type: POOLINGbottom: "conv1"top: "pool1"pooling_param {pool: MAXkernel_size: 3stride: 2}}layers {name: "relu1"type: RELUbottom: "pool1"top: "pool1"}layers {name: "conv2"type: CONVOLUTIONbottom: "pool1"top: "conv2"blobs_lr: 1blobs_lr: 2convolution_param {num_output: 32pad: 2kernel_size: 5stride: 1}}layers {name: "relu2"type: RELUbottom: "conv2"top: "conv2"}layers {name: "pool2"type: POOLINGbottom: "conv2"top: "pool2"pooling_param {pool: AVEkernel_size: 3stride: 2}}layers {name: "conv3"type: CONVOLUTIONbottom: "pool2"top: "conv3"blobs_lr: 1blobs_lr: 2convolution_param {num_output: 64pad: 2kernel_size: 5stride: 1}}layers {name: "relu3"type: RELUbottom: "conv3"top: "conv3"}layers {name: "pool3"type: POOLINGbottom: "conv3"top: "pool3"pooling_param {pool: AVEkernel_size: 3stride: 2}}layers {name: "ip1"type: INNER_PRODUCTbottom: "pool3"top: "ip1"blobs_lr: 1blobs_lr: 2inner_product_param {num_output: 64}}layers {name: "ip2"type: INNER_PRODUCTbottom: "ip1"top: "ip2"blobs_lr: 1blobs_lr: 2inner_product_param {num_output: 10}}layers {name: "prob"type: SOFTMAXbottom: "ip2"top: "prob"}

复制代码

可见，上述模型经过了3个卷积层（conv1, conv2, conv3）(👼)，每个(😟)卷(🐋)积层后面都(🦍)跟着下采样层（pool1, pool2, pool3），之后有两个全连接层(ip1, ip2)，最后一层prob为SOFTMAX分(⏸)类层，是计算概率密度的，这(🍣)里我(🍃)们不需要关心。

下面三张图分别统计了CNN模型各层的参数量、数据量和计算量。

可以(🕞)看出，卷积层的参数量很少，但数据量很大；全连接层刚好相反，参数量较大，但数据量(✈)很少。

通过计算量统计发现conv2计算量(💿)最大，其次是conv3和conv1。全连接层的(✒)计算量相对卷积层较小，但不可忽略(🕷)。其他层（pool1, pool2以及各级relu）由于计算量太小，本设计中没有将其实现为Open CL kernel，而是直接CPU端实现。

综(⤵)上所述，我们重点实(❎)现两个算法：卷积和矩阵乘，分别对应(🗯)卷积层、全连接层的实现。

在(⛑)DE1-SOC上我利用了友晶提供的Open CL BSP，支持C语言开发FPGA。

卷积层计算kernel函数如下：

__attribute__((num_compute_units(4)))__kernelvoid conv(__global float * a, __global float * b, __global float * c, const int M, const int N, const int K){int gx = get_global_id(0);int gy = get_global_id(1);float tmp=0.0f;for(int x = 0; x < K; x ++){for(int y = 0; y < K; y ++){tmp += a[(gx + x) * M + (gy + y)] * b[x * K + y];}}

复制代码

全连(⛹)接层计算采用矩阵乘实现，kernel函数如下：

__attribute__((num_compute_units(4)))__kernelvoid gemm(__global float * a, __global float * b, __global float * c, const int M, const int N, const int K){int gx = get_global_id(0);int gy = get_global_id(1);int sy = get_global_size(1);int sx = get_global_size(0);int s = sx * sy;for(int x = gx; x < M; x += sx){for(int y = gy; y < N; y += sy){float tmp=0.0f;for(int z = 0; z < K; z++){tmp += a[z * M + x] * b[y * K + z];}c[y * M + x] = tmp;}}}

复制代码

编(🌝)译kernel函数需要使用Altera SDK for OpenCL，我用的版本是14.0.0.200，申请了两个月的license。编译(🌵)使用(🐅)命令行aoc，得到*.aocx文件。

Open CL编译输(😣)出报告中给出了资源占用情况：

+--------------------------------------------------------------------+; Estimated Resource Usage Summary ;+----------------------------------------+---------------------------+; Resource + Usage ;+----------------------------------------+---------------------------+; Logic utilization ; 83% ;; Dedicated logic registers ; 46% ;; Memory blocks ; 57% ;; DSP blocks ; 25% ;+----------------------------------------+---------------------------;

复制代码(🤨)

可见，逻辑资源、存储器资源消耗较为明显(🔩)，而DSP资源并(🤴)未用尽，说明还有优化的空间。

编译主程序需(🕒)要使用SoCEDS，我用的版本(🎩)为14.0.2.274，也是命令行方式，在工程目录下执行make，结束后得到可执行文件cnn。

将这两个文件拷贝到SD卡，按照(🚣)前面的博客对板子(👷)进行设置，将CNN的模型、CIFAR10数据也拷贝到SD卡中，板子上电，mount SD卡到/mnt，执行cnn，得(🤑)到输出如下：

<div class="blockcode"><blockquote>Please input the number of images(1~100):100Loading data...OK!Constructing CNN...OK!Begin calculation...Elapsed Time = 141.861 s.Real Label = 3(cat), Calc Label = 3(cat), error count = 0Real Label = 8(ship), Calc Label = 8(ship), error count = 0Real Label = 8(ship), Calc Label = 8(ship), error count = 0Real Label = 0(airplane), Calc Label = 0(airplane), error count = 0Real Label = 6(frog), Calc Label = 6(frog), error count = 0Real Label = 6(frog), Calc Label = 6(frog), error count = 0Real Label = 1(automobile), Calc Label = 1(automobile), error count = 0Real Label = 6(frog), Calc Label = 6(frog), error count = 0Real Label = 3(cat), Calc Label = 3(cat), error count = 0Real Label = 1(automobile), Calc Label = 1(automobile), error count = 0Real Label = 0(airplane), Calc Label = 0(airplane), error count = 0Real Label = 9(truck), Calc Label = 9(truck), error count = 0Real Label = 5(dog), Calc Label = 5(dog), error count = 0Real Label = 7(horse), Calc Label = 7(horse), error count = 0Real Label = 9(truck), Calc Label = 9(truck), error count = 0Real Label = 8(ship), Calc Label = 8(ship), error count = 0Real Label = 5(dog), Calc Label = 5(dog), error count = 0Real Label = 7(horse), Calc Label = 7(horse), error count = 0Real Label = 8(ship), Calc Label = 8(ship), error count = 0Real Label = 6(frog), Calc Label = 6(frog), error count = 0Real Label = 7(horse), Calc Label = 7(horse), error count = 0Real Label = 0(airplane), Calc Label = 2(bird), error count = 1Real Label = 4(deer), Calc Label = 4(deer), error count = 1Real Label = 9(truck), Calc Label = 9(truck), error count = 1Real Label = 5(dog), Calc Label = 4(deer), error count = 2Real Label = 2(bird), Calc Label = 3(cat), error count = 3Real Label = 4(deer), Calc Label = 4(deer), error count = 3Real Label = 0(airplane), Calc Label = 0(airplane), error count = 3Real Label = 9(truck), Calc Label = 9(truck), error count = 3Real Label = 6(frog), Calc Label = 6(frog), error count = 3Real Label = 6(frog), Calc Label = 6(frog), error count = 3Real Label = 5(dog), Calc Label = 5(dog), error count = 3Real Label = 4(deer), Calc Label = 4(deer), error count = 3Real Label = 5(dog), Calc Label = 5(dog), error count = 3Real Label = 9(truck), Calc Label = 9(truck), error count = 3Real Label = 2(bird), Calc Label = 3(cat), error count = 4Real Label = 4(deer), Calc Label = 7(horse), error count = 5Real Label = 1(automobile), Calc Label = 9(truck), error count = 6Real Label = 9(truck), Calc Label = 9(truck), error count = 6Real Label = 5(dog), Calc Label = 5(dog), error count = 6Real Label = 4(deer), Calc Label = 4(deer), error count = 6Real Label = 6(frog), Calc Label = 6(frog), error count = 6Real Label = 5(dog), Calc Label = 5(dog), error count = 6Real Label = 6(frog), Calc Label = 6(frog), error count = 6Real Label = 0(airplane), Calc Label = 0(airplane), error count = 6Real Label = 9(truck), Calc Label = 9(truck), error count = 6Real Label = 3(cat), Calc Label = 5(dog), error count = 7Real Label = 9(truck), Calc Label = 9(truck), error count = 7Real Label = 7(horse), Calc Label = 7(horse), error count = 7Real Label = 6(frog), Calc Label = 6(frog), error count = 7Real Label = 9(truck), Calc Label = 9(truck), error count = 7Real Label = 8(ship), Calc Label = 8(ship), error count = 7Real Label = 0(airplane), Calc Label = 2(bird), error count = 8Real Label = 3(cat), Calc Label = 3(cat), error count = 8Real Label = 8(ship), Calc Label = 8(ship), error count = 8Real Label = 8(ship), Calc Label = 8(ship), error count = 8Real Label = 7(horse), Calc Label = 7(horse), error count = 8Real Label = 7(horse), Calc Label = 7(horse), error count = 8Real Label = 4(deer), Calc Label = 3(cat), error count = 9Real Label = 6(frog), Calc Label = 3(cat), error count = 10Real Label = 7(horse), Calc Label = 7(horse), error count = 10Real Label = 3(cat), Calc Label = 5(dog), error count = 11Real Label = 6(frog), Calc Label = 6(frog), error count = 11Real Label = 3(cat), Calc Label = 3(cat), error count = 11Real Label = 6(frog), Calc Label = 6(frog), error count = 11Real Label = 2(bird), Calc Label = 2(bird), error count = 11Real Label = 1(automobile), Calc Label = 1(automobile), error count = 11Real Label = 2(bird), Calc Label = 2(bird), error count = 11Real Label = 3(cat), Calc Label = 3(cat), error count = 11Real Label = 7(horse), Calc Label = 9(truck), error count = 12Real Label = 2(bird), Calc Label = 2(bird), error count = 12Real Label = 6(frog), Calc Label = 6(frog), error count = 12Real Label = 8(ship), Calc Label = 8(ship), error count = 12Real Label = 8(ship), Calc Label = 8(ship), error count = 12Real Label = 0(airplane), Calc Label = 0(airplane), error count = 12Real Label = 2(bird), Calc Label = 2(bird), error count = 12Real Label = 9(truck), Calc Label = 0(airplane), error count = 13Real Label = 3(cat), Calc Label = 3(cat), error count = 13Real Label = 3(cat), Calc Label = 2(bird), error count = 14Real Label = 8(ship), Calc Label = 8(ship), error count = 14Real Label = 8(ship), Calc Label = 8(ship), error count = 14Real Label = 1(automobile), Calc Label = 1(automobile), error count = 14Real Label = 1(automobile), Calc Label = 1(automobile), error count = 14Real Label = 7(horse), Calc Label = 7(horse), error count = 14Real Label = 2(bird), Calc Label = 2(bird), error count = 14Real Label = 5(dog), Calc Label = 7(horse), error count = 15Real Label = 2(bird), Calc Label = 2(bird), error count = 15Real Label = 7(horse), Calc Label = 7(horse), error count = 15Real Label = 8(ship), Calc Label = 8(ship), error count = 15Real Label = 9(truck), Calc Label = 9(truck), error count = 15Real Label = 0(airplane), Calc Label = 0(airplane), error count = 15Real Label = 3(cat), Calc Label = 4(deer), error count = 16Real Label = 8(ship), Calc Label = 8(ship), error count = 16Real Label = 6(frog), Calc Label = 6(frog), error count = 16Real Label = 4(deer), Calc Label = 4(deer), error count = 16Real Label = 6(frog), Calc Label = 6(frog), error count = 16Real Label = 6(frog), Calc Label = 6(frog), error count = 16Real Label = 0(airplane), Calc Label = 2(bird), error count = 17Real Label = 0(airplane), Calc Label = 0(airplane), error count = 17Real Label = 7(horse), Calc Label = 7(horse), error count = 17Classify Score = 83 %.

上面的执行流程是这样的，首先输入测试样本数目（1到100），由于DE1板子(📈)FPGA端SDRAM容量较小，难以加载全部测试数据（10000张图片），故每次最多装入100张图片(🙄)。之后载入数据到HPS内存，然(😡)后开(🕑)始构建CNN模(💺)型，构建过(✔)程中也实现了Open CL的初始化。构建完毕，将输入图像(☝)依次通过CNN，得到一系列分类结果，与标(🚷)签进行对比，统计错误分(🥕)类个数，计算分类准(⛴)确率。

经过测试，分类准确率达到83%，与Caffe测试结果一致。

经(🍹)过以上测试，可以得到结论：

（1）使用Open CL可以很方便地移植高级语言编写的算法；

（2）CNN在移植过程中需要考虑实际硬件，定制(🍶)合适的模型和数据(👐)；

（3）Cyclone 5逻辑资源较少（85K，Open CL kernel占用了83%），如果希望进一步(🍺)提高计算速度(🚗)，一方面可以选用高性能器件（如Stratix V、Arria 10），另一方面可以使用RTL自己搭建计算系(👓)统。

以上图文内容均是EEWORLD论坛网友zhaoyongke原创，在此感谢。

欢迎微博@EEWORLD

如果你也写过此类原创干货请关注微信订(👦)阅号（ID：eeworldbbs，将你(🌝)的原创发至：bbs_service@eeworld.com.cn，一经入选，我们将帮你(🐰)登上头条！

与更(😞)多行业内网友进行交流请登陆EEWORLD论坛。

【看黄色aa大片的相关新闻】