1.分类任务

Fashion-mnist分类任务：针对

Fashion-MNIST

数据集，设计、搭建、训练机器学习模型，能够尽可能准确地分辨出测试数据地标签。

前言

1.1 利用VGG模型

import os
import sys
import time
import torch
from torch import nn, optim
import torch.nn.functional as F
import torchvision
import matplotlib.pyplot as plt
from torchvision import transforms

def load_data_fashion_mnist(batch_size, resize=None, root='./FashionMNIST'):
    """Download the fashion mnist dataset and then load into memory."""
    trans = []
    if resize:
        trans.append(torchvision.transforms.Resize(size=resize))
    trans.append(torchvision.transforms.ToTensor())
    trans.append(torchvision.transforms.Normalize((0.1307,), (0.3081,)))
    
    transform = torchvision.transforms.Compose(trans)
    mnist_train = torchvision.datasets.FashionMNIST(root=root, train=True, download=True, transform=transform)
    mnist_test = torchvision.datasets.FashionMNIST(root=root, train=False, download=True, transform=transform)

    train_iter = torch.utils.data.DataLoader(mnist_train, batch_size=batch_size, shuffle=True, num_workers=2)
    test_iter = torch.utils.data.DataLoader(mnist_test, batch_size=batch_size, shuffle=False, num_workers=2)

    return train_iter, test_iter



print('训练...')
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

batch_size = 64  
train_iter, test_iter = load_data_fashion_mnist(batch_size)




def vgg_block(num_convs, in_channels, out_channels): #卷积层个数，输入通道数，输出通道数
    blk = []
    for i in range(num_convs):
        if i == 0:
            blk.append(nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1))
        else:
            blk.append(nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1))
        blk.append(nn.ReLU())
    blk.append(nn.MaxPool2d(kernel_size=2, stride=2)) # 这里会使宽高减半 b*64*14*14 
    return nn.Sequential(*blk)
conv_arch = ((1, 1, 64), (1, 64, 128))
# 经过28/4=7
fc_features = 128 * 7 * 7 # c * w * h
fc_hidden_units = 2048 # 任意
def vgg(conv_arch, fc_features, fc_hidden_units=4096):
    net = nn.Sequential()
    # 卷积层部分
    for i, (num_convs, in_channels, out_channels) in enumerate(conv_arch):
        # 每经过一个vgg_block都会使宽高减半
        net.add_module("vgg_block_" + str(i+1), vgg_block(num_convs, in_channels, out_channels))
    # 全连接层部分
    net.add_module("fc", nn.Sequential(nn.Flatten(),
                                 nn.Linear(fc_features, fc_hidden_units),
                                 nn.ReLU(),
                                 nn.Dropout(0.5),
                                 nn.Linear(fc_hidden_units, fc_hidden_units),
                                 nn.ReLU(),
                                 nn.Dropout(0.5),
                                 nn.Linear(fc_hidden_units, 10)
                                ))
    return net
net = vgg(conv_arch, fc_features, fc_hidden_units)
print(net)


def evaluate_accuracy(data_iter, net, device=None):
    if device is None and isinstance(net, torch.nn.Module):
        # 如果没指定device就使用net的device
        device = list(net.parameters())[0].device
    net.eval() 
    acc_sum, n = 0.0, 0
    with torch.no_grad():
        for X, y in data_iter:
            acc_sum += (net(X.to(device)).argmax(dim=1) == y.to(device)).float().sum().cpu().item()
            n += y.shape[0]
    net.train() # 改回训练模式
    return acc_sum / n

def train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs):
    net = net.to(device)
    print("training on ", device)
    loss = torch.nn.CrossEntropyLoss()
    best_test_acc =0
    for epoch in range(num_epochs):
        train_l_sum, train_acc_sum, n, batch_count, start = 0.0, 0.0, 0, 0, time.time()
        for X, y in train_iter:
            X = X.to(device)
            y = y.to(device)
            y_hat = net(X)
            l = loss(y_hat, y)
#             l2= torch.tensor(0).float().cpu()
#             l = loss(y_hat, y).float().cpu()
 
           
#             for param in net.parameters():
#                  l2 += torch.norm(param, 2).float().cpu()
#             l=(l+l2).cpu()
            optimizer.zero_grad()
            l.backward()
            optimizer.step()
            train_l_sum += l.cpu().item()
            train_acc_sum += (y_hat.argmax(dim=1) == y).sum().cpu().item()
            n += y.shape[0]
            batch_count += 1
        test_acc = evaluate_accuracy(test_iter, net,device)
        print('epoch %d, loss %.4f, train acc %.3f, test acc %.3f, time %.1f sec'
              % (epoch + 1, train_l_sum / batch_count, train_acc_sum / n, test_acc, time.time() - start))
              
        if test_acc>best_test_acc:
            print('find best! save at model/best.pth')
            best_test_acc = test_acc
            torch.save(net.state_dict(), 'model/best.pth')



lr, num_epochs = 0.001, 10
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)            

结果


利用vgg 设计的网络差不多测试结果准确率能够达到0.92。

1.2 使用Resnet

import os
import sys
import time
import torch
from torch import nn, optim
import torch.nn.functional as F
import torchvision
import matplotlib.pyplot as plt
from torchvision import transforms


def load_data_fashion_mnist(batch_size, resize=None, root='./FashionMNIST'):
    """Download the fashion mnist dataset and then load into memory."""
    trans = []
    if resize:
        trans.append(torchvision.transforms.Resize(size=resize))
    trans.append(torchvision.transforms.ToTensor())
    trans.append(torchvision.transforms.Normalize((0.1307,), (0.3081,)))
    
    transform = torchvision.transforms.Compose(trans)
    mnist_train = torchvision.datasets.FashionMNIST(root=root, train=True, download=True, transform=transform)
    mnist_test = torchvision.datasets.FashionMNIST(root=root, train=False, download=True, transform=transform)

    train_iter = torch.utils.data.DataLoader(mnist_train, batch_size=batch_size, shuffle=True, num_workers=2)
    test_iter = torch.utils.data.DataLoader(mnist_test, batch_size=batch_size, shuffle=False, num_workers=2)

    return train_iter, test_iter


print('训练...')
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

batch_size = 64  
train_iter, test_iter = load_data_fashion_mnist(batch_size)



class Residual(nn.Module):  # 本类已保存在d2lzh_pytorch包中方便以后使用
    #可以设定输出通道数、是否使用额外的1x1卷积层来修改通道数以及卷积层的步幅。
    def __init__(self, in_channels, out_channels, use_1x1conv=False, stride=1):
        super(Residual, self).__init__()
        self.conv1 = nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1, stride=stride)
        self.conv2 = nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1)
        if use_1x1conv:
            self.conv3 = nn.Conv2d(in_channels, out_channels, kernel_size=1, stride=stride)
        else:
            self.conv3 = None
        self.bn1 = nn.BatchNorm2d(out_channels)
        self.bn2 = nn.BatchNorm2d(out_channels)

    def forward(self, X):
        Y = F.relu(self.bn1(self.conv1(X)))
        Y = self.bn2(self.conv2(Y))
        if self.conv3:
            X = self.conv3(X)
        return F.relu(Y + X)

def resnet_block(in_channels, out_channels, num_residuals, first_block=False):
    if first_block:
        assert in_channels == out_channels # 第一个模块的通道数同输入通道数一致
    blk = []
    for i in range(num_residuals):
        if i == 0 and not first_block:
            blk.append(Residual(in_channels, out_channels, use_1x1conv=True, stride=2))
        else:
            blk.append(Residual(out_channels, out_channels))
    return nn.Sequential(*blk)


net = nn.Sequential(
        nn.Conv2d(1, 32, kernel_size=3, stride=1, padding=1),
        nn.BatchNorm2d(32), 
        nn.ReLU(),
        nn.MaxPool2d(kernel_size=2, stride=2))
net.add_module("resnet_block1", resnet_block(32, 32, 2, first_block=True))
net.add_module("resnet_block2", resnet_block(32, 64, 2))
net.add_module("resnet_block3", resnet_block(64, 128, 2))
net.add_module("resnet_block4", resnet_block(128, 256, 2))

class GlobalAvgPool2d(nn.Module):
    # 全局平均池化层可通过将池化窗口形状设置成输入的高和宽实现
    def __init__(self):
        super(GlobalAvgPool2d, self).__init__()
    def forward(self, x):
        return F.avg_pool2d(x, kernel_size=x.size()[2:])
class FlattenLayer(torch.nn.Module):
    def __init__(self):
        super(FlattenLayer, self).__init__()
    def forward(self, x): # x shape: (batch, *, *, ...)
        return x.view(x.shape[0], -1)
    
net.add_module("global_avg_pool", GlobalAvgPool2d()) # GlobalAvgPool2d的输出: (Batch, 512, 1, 1)
net.add_module("fc", nn.Sequential(FlattenLayer(), nn.Linear(256, 10))) 



def train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs):
    net = net.to(device)
    print("training on ", device)
    loss = torch.nn.CrossEntropyLoss()
    best_test_acc =0
    for epoch in range(num_epochs):
        train_l_sum, train_acc_sum, n, batch_count, start = 0.0, 0.0, 0, 0, time.time()
        for X, y in train_iter:
            X = X.to(device)
            y = y.to(device)
            y_hat = net(X)
            l = loss(y_hat, y)
            optimizer.zero_grad()
            l.backward()
            optimizer.step()
            train_l_sum += l.cpu().item()
            train_acc_sum += (y_hat.argmax(dim=1) == y).sum().cpu().item()
            n += y.shape[0]
            batch_count += 1
        test_acc = evaluate_accuracy(test_iter, net,device)
        print('epoch %d, loss %.4f, train acc %.3f, test acc %.3f, time %.1f sec'
              % (epoch + 1, train_l_sum / batch_count, train_acc_sum / n, test_acc, time.time() - start))
              
        if test_acc>best_test_acc:
            print('find best! save at model/best.pth')
            best_test_acc = test_acc
            torch.save(net.state_dict(), 'model/best2.pth')
    




lr, num_epochs = 0.001, 10
optimizer = torch.optim.Adm(net.parameters(), lr=lr)
train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)        

结果显示


训练10轮差不多能达到0.925.



参考：https://github.com/monkeyDemon/Learn_Dive-into-DL-PyTorch