第八课：ShuffleNet v1、ShuffleNet v2学习

前言

随着人工智能的不断发展，机器学习这门技术也越来越重要，很多人都开启了学习机器学习，本文就介绍了机器学习的基础内容。来源于哔哩哔哩博主“霹雳吧啦Wz”，博主学习作为笔记记录，欢迎大家一起讨论学习交流。

一、搭建ShuffleNet网络

用ShuffleNet v2学习,轮次训练11轮次（代码），个epoch大概7-8分钟。

二、代码部分

1.module.py----定义ShuffleNet的网络结构

代码如下（示例）：

from typing import List, Callable

import torch
from torch import Tensor
import torch.nn as nn


def channel_shuffle(x: Tensor, groups: int) -> Tensor:#这个部分实现channe shuffle思想，见博客图

    batch_size, num_channels, height, width = x.size()
    channels_per_group = num_channels // groups

    # reshape
    # [batch_size, num_channels, height, width] -> [batch_size, groups, channels_per_group, height, width]
    x = x.view(batch_size, groups, channels_per_group, height, width)

    x = torch.transpose(x, 1, 2).contiguous()

    # flatten
    x = x.view(batch_size, -1, height, width)

    return x


class InvertedResidual(nn.Module):#Shuffle Net的block模块inverted residual
    def __init__(self, input_c: int, output_c: int, stride: int):
        super(InvertedResidual, self).__init__()

        if stride not in [1, 2]:#判断步距是否是1和2
            raise ValueError("illegal stride value.")
        self.stride = stride

        assert output_c % 2 == 0
        branch_features = output_c // 2
        # 当stride为1时，input_channel应该是branch_features的两倍
        # python中 '<<' 是位运算，可理解为计算×2的快速方法
        assert (self.stride != 1) or (input_c == branch_features << 1)

        if self.stride == 2:
            self.branch1 = nn.Sequential(
                self.depthwise_conv(input_c, input_c, kernel_s=3, stride=self.stride, padding=1),
                nn.BatchNorm2d(input_c),
                nn.Conv2d(input_c, branch_features, kernel_size=1, stride=1, padding=0, bias=False),
                nn.BatchNorm2d(branch_features),
                nn.ReLU(inplace=True)
            )
        else:
            self.branch1 = nn.Sequential()

        self.branch2 = nn.Sequential(
            nn.Conv2d(input_c if self.stride > 1 else branch_features, branch_features, kernel_size=1,
                      stride=1, padding=0, bias=False),
            nn.BatchNorm2d(branch_features),
            nn.ReLU(inplace=True),
            self.depthwise_conv(branch_features, branch_features, kernel_s=3, stride=self.stride, padding=1),
            nn.BatchNorm2d(branch_features),
            nn.Conv2d(branch_features, branch_features, kernel_size=1, stride=1, padding=0, bias=False),
            nn.BatchNorm2d(branch_features),
            nn.ReLU(inplace=True)
        )

    @staticmethod
    def depthwise_conv(input_c: int,
                       output_c: int,
                       kernel_s: int,
                       stride: int = 1,
                       padding: int = 0,
                       bias: bool = False) -> nn.Conv2d:
        return nn.Conv2d(in_channels=input_c, out_channels=output_c, kernel_size=kernel_s,
                         stride=stride, padding=padding, bias=bias, groups=input_c)

    def forward(self, x: Tensor) -> Tensor:
        if self.stride == 1:
            x1, x2 = x.chunk(2, dim=1)
            out = torch.cat((x1, self.branch2(x2)), dim=1)
        else:
            out = torch.cat((self.branch1(x), self.branch2(x)), dim=1)

        out = channel_shuffle(out, 2)

        return out


class ShuffleNetV2(nn.Module):#构建ShuffleNetV2
    def __init__(self,
                 stages_repeats: List[int],
                 stages_out_channels: List[int],
                 num_classes: int = 1000,
                 inverted_residual: Callable[..., nn.Module] = InvertedResidual):
        super(ShuffleNetV2, self).__init__()

        if len(stages_repeats) != 3:
            raise ValueError("expected stages_repeats as list of 3 positive ints")
        if len(stages_out_channels) != 5:
            raise ValueError("expected stages_out_channels as list of 5 positive ints")
        self._stage_out_channels = stages_out_channels

        # input RGB image
        input_channels = 3
        output_channels = self._stage_out_channels[0]

        self.conv1 = nn.Sequential(
            nn.Conv2d(input_channels, output_channels, kernel_size=3, stride=2, padding=1, bias=False),
            nn.BatchNorm2d(output_channels),
            nn.ReLU(inplace=True)
        )
        input_channels = output_channels

        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, padding=1)

        # Static annotations for mypy
        self.stage2: nn.Sequential
        self.stage3: nn.Sequential
        self.stage4: nn.Sequential

        stage_names = ["stage{}".format(i) for i in [2, 3, 4]]
        for name, repeats, output_channels in zip(stage_names, stages_repeats,
                                                  self._stage_out_channels[1:]):
            seq = [inverted_residual(input_channels, output_channels, 2)]# 步距为2
            for i in range(repeats - 1):
                seq.append(inverted_residual(output_channels, output_channels, 1))# 步距为1
            setattr(self, name, nn.Sequential(*seq))
            input_channels = output_channels

        output_channels = self._stage_out_channels[-1]
        self.conv5 = nn.Sequential(
            nn.Conv2d(input_channels, output_channels, kernel_size=1, stride=1, padding=0, bias=False),
            nn.BatchNorm2d(output_channels),
            nn.ReLU(inplace=True)
        )

        self.fc = nn.Linear(output_channels, num_classes)

    def _forward_impl(self, x: Tensor) -> Tensor:
        # See note [TorchScript super()]
        x = self.conv1(x)
        x = self.maxpool(x)
        x = self.stage2(x)
        x = self.stage3(x)
        x = self.stage4(x)
        x = self.conv5(x)
        x = x.mean([2, 3])  # global pool 2，3对应高度和宽度两个维度 只剩下batch和channel维度
        x = self.fc(x)
        return x

    def forward(self, x: Tensor) -> Tensor:
        return self._forward_impl(x)


def shufflenet_v2_x0_5(num_classes=1000):
    """
    Constructs a ShuffleNetV2 with 0.5x output channels, as described in
    `"ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design"
    <https://arxiv.org/abs/1807.11164>`.
    weight: https://download.pytorch.org/models/shufflenetv2_x0.5-f707e7126e.pth

    :param num_classes:
    :return:
    """
    model = ShuffleNetV2(stages_repeats=[4, 8, 4],
                         stages_out_channels=[24, 48, 96, 192, 1024],
                         num_classes=num_classes)

    return model


def shufflenet_v2_x1_0(num_classes=1000):
    """
    Constructs a ShuffleNetV2 with 1.0x output channels, as described in
    `"ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design"
    <https://arxiv.org/abs/1807.11164>`.
    weight: https://download.pytorch.org/models/shufflenetv2_x1-5666bf0f80.pth

    :param num_classes:
    :return:
    """
    model = ShuffleNetV2(stages_repeats=[4, 8, 4],
                         stages_out_channels=[24, 116, 232, 464, 1024],
                         num_classes=num_classes)

    return model


def shufflenet_v2_x1_5(num_classes=1000):
    """
    Constructs a ShuffleNetV2 with 1.0x output channels, as described in
    `"ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design"
    <https://arxiv.org/abs/1807.11164>`.
    weight: https://download.pytorch.org/models/shufflenetv2_x1_5-3c479a10.pth

    :param num_classes:
    :return:
    """
    model = ShuffleNetV2(stages_repeats=[4, 8, 4],
                         stages_out_channels=[24, 176, 352, 704, 1024],
                         num_classes=num_classes)

    return model


def shufflenet_v2_x2_0(num_classes=1000):
    """
    Constructs a ShuffleNetV2 with 1.0x output channels, as described in
    `"ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design"
    <https://arxiv.org/abs/1807.11164>`.
    weight: https://download.pytorch.org/models/shufflenetv2_x2_0-8be3c8ee.pth

    :param num_classes:
    :return:
    """
    model = ShuffleNetV2(stages_repeats=[4, 8, 4],
                         stages_out_channels=[24, 244, 488, 976, 2048],
                         num_classes=num_classes)

    return model

2.train.py----加载数据集并进行训练，训练集计算loss，测试集计算accuracy，保存训练好的网络参数

代码如下（示例）：

import os
import math
import argparse

import torch
import torch.optim as optim
from torch.utils.tensorboard import SummaryWriter
from torchvision import transforms
import torch.optim.lr_scheduler as lr_scheduler

from model import shufflenet_v2_x1_0
from my_dataset import MyDataSet
from utils import read_split_data, train_one_epoch, evaluate


def main(args):
    device = torch.device(args.device if torch.cuda.is_available() else "cpu")

    print(args)
    print('Start Tensorboard with "tensorboard --logdir=runs", view at http://localhost:6006/')
    tb_writer = SummaryWriter()
    if os.path.exists("./weights") is False:
        os.makedirs("./weights")

    train_images_path, train_images_label, val_images_path, val_images_label = read_split_data(args.data_path)

    data_transform = {
        "train": transforms.Compose([transforms.RandomResizedCrop(224),
                                     transforms.RandomHorizontalFlip(),
                                     transforms.ToTensor(),
                                     transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])]),
        "val": transforms.Compose([transforms.Resize(256),
                                   transforms.CenterCrop(224),
                                   transforms.ToTensor(),
                                   transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])}

    # 实例化训练数据集
    train_dataset = MyDataSet(images_path=train_images_path,
                              images_class=train_images_label,
                              transform=data_transform["train"])

    # 实例化验证数据集
    val_dataset = MyDataSet(images_path=val_images_path,
                            images_class=val_images_label,
                            transform=data_transform["val"])

    batch_size = args.batch_size
    nw = min([os.cpu_count(), batch_size if batch_size > 1 else 0, 8])  # number of workers
    print('Using {} dataloader workers every process'.format(nw))
    train_loader = torch.utils.data.DataLoader(train_dataset,
                                               batch_size=batch_size,
                                               shuffle=True,
                                               pin_memory=True,
                                               num_workers=nw,
                                               collate_fn=train_dataset.collate_fn)

    val_loader = torch.utils.data.DataLoader(val_dataset,
                                             batch_size=batch_size,
                                             shuffle=False,
                                             pin_memory=True,
                                             num_workers=nw,
                                             collate_fn=val_dataset.collate_fn)

    # 如果存在预训练权重则载入
    model = shufflenet_v2_x1_0(num_classes=args.num_classes).to(device)
    if args.weights != "":
        if os.path.exists(args.weights):
            weights_dict = torch.load(args.weights, map_location=device)
            load_weights_dict = {k: v for k, v in weights_dict.items()
                                 if model.state_dict()[k].numel() == v.numel()}
            print(model.load_state_dict(load_weights_dict, strict=False))
        else:
            raise FileNotFoundError("not found weights file: {}".format(args.weights))

    # 是否冻结权重
    if args.freeze_layers:
        for name, para in model.named_parameters():
            # 除最后的全连接层外，其他权重全部冻结
            if "fc" not in name:
                para.requires_grad_(False)

    pg = [p for p in model.parameters() if p.requires_grad]
    optimizer = optim.SGD(pg, lr=args.lr, momentum=0.9, weight_decay=4E-5)
    # Scheduler https://arxiv.org/pdf/1812.01187.pdf
    lf = lambda x: ((1 + math.cos(x * math.pi / args.epochs)) / 2) * (1 - args.lrf) + args.lrf  # cosine
    scheduler = lr_scheduler.LambdaLR(optimizer, lr_lambda=lf)

    for epoch in range(args.epochs):
        # train
        mean_loss = train_one_epoch(model=model,
                                    optimizer=optimizer,
                                    data_loader=train_loader,
                                    device=device,
                                    epoch=epoch)

        scheduler.step()

        # validate
        acc = evaluate(model=model,
                       data_loader=val_loader,
                       device=device)

        print("[epoch {}] accuracy: {}".format(epoch, round(acc, 3)))
        tags = ["loss", "accuracy", "learning_rate"]
        tb_writer.add_scalar(tags[0], mean_loss, epoch)
        tb_writer.add_scalar(tags[1], acc, epoch)
        tb_writer.add_scalar(tags[2], optimizer.param_groups[0]["lr"], epoch)

        torch.save(model.state_dict(), "./weights/model-{}.pth".format(epoch))


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('--num_classes', type=int, default=5)
    parser.add_argument('--epochs', type=int, default=30)
    parser.add_argument('--batch-size', type=int, default=16)
    parser.add_argument('--lr', type=float, default=0.01)
    parser.add_argument('--lrf', type=float, default=0.1)

    # 数据集所在根目录
    # https://storage.googleapis.com/download.tensorflow.org/example_images/flower_photos.tgz
    parser.add_argument('--data-path', type=str,
                        default="G:/深度学习/deep-learning-for-image-processing-master/data_set/flower_data/flower_photos")

    # shufflenetv2_x1.0 官方权重下载地址
    # https://download.pytorch.org/models/shufflenetv2_x1-5666bf0f80.pth
    parser.add_argument('--weights', type=str, default='./shufflenetv2_x1.pth',
                        help='initial weights path')
    parser.add_argument('--freeze-layers', type=bool, default=False)#False训练所有层结构 True训练全连接层
    parser.add_argument('--device', default='cuda:0', help='device id (i.e. 0 or 0,1 or cpu)')

    opt = parser.parse_args()

    main(opt)

3.predict.py——得到训练好的网络参数后，用自己找的图像进行分类测试

import os
import json

import torch
from PIL import Image
from torchvision import transforms
import matplotlib.pyplot as plt

from model import shufflenet_v2_x1_0


def main():
    device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

    data_transform = transforms.Compose(
        [transforms.Resize(256),
         transforms.CenterCrop(224),
         transforms.ToTensor(),
         transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])

    # load image
    img_path = "../tulip.jpg"
    assert os.path.exists(img_path), "file: '{}' dose not exist.".format(img_path)
    img = Image.open(img_path)
    plt.imshow(img)
    # [N, C, H, W]
    img = data_transform(img)
    # expand batch dimension
    img = torch.unsqueeze(img, dim=0)

    # read class_indict
    json_path = './class_indices.json'
    assert os.path.exists(json_path), "file: '{}' dose not exist.".format(json_path)

    with open(json_path, "r") as f:
        class_indict = json.load(f)

    # create model
    model = shufflenet_v2_x1_0(num_classes=5).to(device)
    # load model weights
    model_weight_path = "./weights/model-29.pth"
    model.load_state_dict(torch.load(model_weight_path, map_location=device))
    model.eval()
    with torch.no_grad():
        # predict class
        output = torch.squeeze(model(img.to(device))).cpu()
        predict = torch.softmax(output, dim=0)
        predict_cla = torch.argmax(predict).numpy()

    print_res = "class: {}   prob: {:.3}".format(class_indict[str(predict_cla)],
                                                 predict[predict_cla].numpy())
    plt.title(print_res)
    for i in range(len(predict)):
        print("class: {:10}   prob: {:.3}".format(class_indict[str(i)],
                                                  predict[i].numpy()))
    plt.show()


if __name__ == '__main__':
    main()

三、ShuffleNet v1

channel shuffle:用组卷积和DW卷积

对比AlexNet,ShuffleNet 0.5容错率差不多，推理时间15.2ms,AlexNet是10ms,说明ShffleNet非常的轻量级。

普通的组卷积是可以减少参数，但是不能使不同组卷积之间进行信息交流，所以提出了channel shuffle的概念。对于组卷积分成三组，每一组又分成三小组，把每一组对应位置索引的小组，组合在一起，构成一个新的组，这个组就包含了之前分组卷积没有的的各个组卷积的信息，实现了各组的channnel信息交流。

实验表明：ResNeXt的PW1x1的普通卷积理论上占据内存的百分之93.4.

因为前面说了1x1的卷积占据内存比较大，所以将1x1卷积换成了组卷积。中间是针对stride=1的情况，右边是针对stride=2的情况。右图是步距为2，捷径是通过池化进行，之前是卷积，最终再进行concat相加。

大部分采用g=3的情况，通过输入224x224的图像，经过第一个卷积Conv1,卷积核大小是3x3,步距为2，重复一次，输出特征矩阵channel是24（卷积核1使用24个卷积核），然后再经过stage2,stage3,stage4…经过stage2的时候，对步距为2的block堆叠1次，输出矩阵channel240，对步距为1的block堆叠3次，输出矩阵channel240。

对于每一个stage的第一个block步距都是2.对于下一个stage的输出矩阵的Chanel通道数翻倍。从表格里面也可以看出来。对于Resnet而言，第一个卷积核和第二个卷积核的个数是输出矩阵个数通道数的1/4.

对于ShffleNet，GConv和DConv是同样的道理，也是输出矩阵的通道数的1/4。

对于stage2，第一个卷积不是使用组卷积。

ResNeXt在ResNet基础上，在第二个3x3卷积上采用了组卷积。DW卷积是组卷积的一个特殊情况：当组卷积的分组数g=输入特征矩阵的深度Cin，输出矩阵深度=输入矩阵深度时候，分组卷积变成了DW卷积。

总结：ShuffleNet参数理论上计算量最小。主要思想是在Group 卷积后面加上一个channel shuffle模块，对于block而言，我们将1x1卷积替换成1x1组卷积。

四、ShuffleNet v2

Flops:浮点运算数

就是不能仅仅看Flops，速度才是最直接的因素。MAC:内存访问时间成本。还需要考虑。并行等级。相同Flops时候，要看并行等级的高低。另外，相同的Flops在不同的平台也不一样

基于这个观点 作者提出了一些措施。

G1:当卷积层的输入特征矩阵与输出矩阵的channel相等时候MAC最小（Flops不变）

随着比值的增加，推理速度越来越慢。

G2:当GConv的groups增大时候（保持Flos不变），MAC也会增大。

hwc1:输入 hwc2:输出

随着g的增加，cpu和gpu都下降了。

G3:网络设计的碎片化程度（分支）越高，速度越慢。（右上角实验也可以说明）

碎片化结构能够提高模型的准确率，但是会降低效率。

G4:Eliment-wise操作带来的影响不可忽视

Relu、AddTensor、AddBias等，Flops小，但是MAC较大。

总结：

1.平衡的convolutions:输入与输出特征矩阵channel尽可能为1

2.注意组卷积的groups:过多增加计算成本

3.减少分支

4.减少使用element-wise operations

a是步距为1，b是步距为2的ShuffleNet的网络结构。

G3:减少分支

G1:平衡的convolutions:输入与输出特征矩阵channel尽可能为1

G2:注意组卷积的groups:过多增加计算成本

G1:平衡的convolutions:输入与输出特征矩阵channel尽可能为1

add在Shuffle NetV2中不存在，Relu仅仅在一个分支进行。也可以将Concat、Channel Shuffle、Channel Split合并。

对于步距为2，channel split不存在，深度翻倍。

c=c/2

和Shuffle Netv1 唯一不同，多了一个1x1卷积层，即Conv5.(对比Shuffle Netv1）

总结：

1.stride=1进行channel Split，分成两个分支，主分支由原来的1x1组卷积变成普通1x1卷积，Relu变成主分支。Add换成Concat操作。以及channel shuffle

2.stride=2,捷径分支由池化变成DW＋Conv。

最后看看ShuffleNetV2的一些指标：

错误率比较低。

关于代码部分 def channel_shuffle部分：输入图片格式是batch_size,channels,height,width.这里假设上图输入是6个channel，把它分成三个组groups，每组就有两个channels_groups。通过x.view函数即可实现从左图到右图的方式。

通过torch.transpose函数将维度1（groups=3）和维度2（channels_groups=2）进行对调.

再通过x.view函数还原成batchsize,channels,height和width.对比最原始的图片（上图）即实现了ShuffleNet的Channel ShuffleNet思想。