import torch 
from torch import nn

def corr2d(X, K):  # 本函数已保存在d2lzh_pytorch包中方便以后使用
    h, w = K.shape
    # 在卷积运算之后结果的形状
    Y = torch.zeros((X.shape[0] - h + 1, X.shape[1] - w + 1))
    for i in range(Y.shape[0]):
        for j in range(Y.shape[1]):
            #指定区域与卷积核进行运算
            Y[i, j] = (X[i: i + h, j: j + w] * K).sum()
    return Y

X = torch.tensor([[0, 1, 2], [3, 4, 5], [6, 7, 8]])
K = torch.tensor([[0, 1], [2, 3]])
corr2d(X, K)

class Conv2D(nn.Module):
    def __init__(self, kernel_size):
        super(Conv2D, self).__init__()
        self.weight = nn.Parameter(torch.randn(kernel_size))
        # 标准正态分布，均值为0，方差为1
        self.bias = nn.Parameter(torch.randn(1))

    def forward(self, x):
        return corr2d(x, self.weight) + self.bias

X = torch.ones(6, 8)
X[:, 2:6] = 0
X

K = torch.tensor([[1, -1]])

Y = corr2d(X, K)
Y

# 构造一个核数组形状是(1, 2)的二维卷积层
conv2d = Conv2D(kernel_size=(1, 2))

step = 20
lr = 0.01
for i in range(step):
    Y_hat = conv2d(X)
    l = ((Y_hat - Y) ** 2).sum()
    l.backward()
    
    # 梯度下降
    conv2d.weight.data -= lr * conv2d.weight.grad
    conv2d.bias.data -= lr * conv2d.bias.grad
    
    # 梯度清0
    conv2d.weight.grad.fill_(0)
    conv2d.bias.grad.fill_(0)
    if (i + 1) % 5 == 0:
        print('Step %d, loss %.3f' % (i + 1, l.item()))

print("weight: ", conv2d.weight.data)
print("bias: ", conv2d.bias.data)

import torch
from torch import nn

# 定义一个函数来计算卷积层。它对输入和输出做相应的升维和降维
def comp_conv2d(conv2d, X):
    # (1, 1)代表批量大小和通道数（“多输入通道和多输出通道”一节将介绍）均为1
    # # 此步将X由[8,8]变为[1,1,8,8]
    X = X.view((1, 1) + X.shape)
    Y = conv2d(X)
    return Y.view(Y.shape[2:])  # 排除不关心的前两维：批量和通道

# padding=1注意这里是两侧分别填充1行或列，所以在两侧一共填充2行或列
# 定义输入通道为数据的通道数，与上一层的输出通道数相同
# 定义的输出通道数为结果的通道数，留到下一层使用
conv2d = nn.Conv2d(in_channels=1, out_channels=1, kernel_size=3, padding=1)

X = torch.rand(8, 8)
comp_conv2d(conv2d, X).shape

# 使用高为5、宽为3的卷积核。在高和宽两侧的填充数分别为2和1
conv2d = nn.Conv2d(in_channels=1, out_channels=1, kernel_size=(5, 3), padding=(2, 1))
comp_conv2d(conv2d, X).shape

conv2d = nn.Conv2d(1, 1, kernel_size=3, padding=1, stride=2)
comp_conv2d(conv2d, X).shape

conv2d = nn.Conv2d(1, 1, kernel_size=(3, 5), padding=(0, 1), stride=(3, 4))
comp_conv2d(conv2d, X).shape

import torch
from torch import nn
import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l

def corr2d_multi_in(X, K):
    # 先计算了第0维得到了形状
    # 沿着X和K的第0维（通道维）分别计算再相加
    res = d2l.corr2d(X[0, :, :], K[0, :, :])
    # 进行其余通道的二维卷积运算
    for i in range(1, X.shape[0]):
        res += d2l.corr2d(X[i, :, :], K[i, :, :])
    return res

# [2, 3, 3]
X = torch.tensor([[[0, 1, 2], [3, 4, 5], [6, 7, 8]],
              [[1, 2, 3], [4, 5, 6], [7, 8, 9]]])
# [2, 2, 2]
K = torch.tensor([[[0, 1], [2, 3]], [[1, 2], [3, 4]]])

corr2d_multi_in(X, K)

def corr2d_multi_in_out(X, K):
    # 对K的第0维遍历，每次同输入X做互相关计算。所有结果使用stack函数合并在一起
    # 前两个k为小k，最后一个K为参数大K
    return torch.stack([corr2d_multi_in(X, k) for k in K])

K = torch.stack([K, K + 1, K + 2])
K.shape # torch.Size([3, 2, 2, 2])

corr2d_multi_in_out(X, K)

# 因为相当于全连接层，所以将其转为二维矩阵进行运算，结果重新转回相应形状
def corr2d_multi_in_out_1x1(X, K):
    c_i, h, w = X.shape
    c_o = K.shape[0]
    X = X.view(c_i, h * w)
    K = K.view(c_o, c_i)
    Y = torch.mm(K, X)  # 全连接层的矩阵乘法
    return Y.view(c_o, h, w)

X = torch.rand(3, 3, 3)
K = torch.rand(2, 3, 1, 1)
# 全连接计算的结果
Y1 = corr2d_multi_in_out_1x1(X, K)
# 卷积计算的结果
Y2 = corr2d_multi_in_out(X, K)
# 证明两种是一样的
(Y1 - Y2).norm().item() < 1e-6

import torch
from torch import nn

def pool2d(X, pool_size, mode='max'):
    # 若是没有指定数据的类型，可以使用这种办法
    X = X.float()
    p_h, p_w = pool_size
    Y = torch.zeros(X.shape[0] - p_h + 1, X.shape[1] - p_w + 1)
    for i in range(Y.shape[0]):
        for j in range(Y.shape[1]):
            if mode == 'max':
                # 池化尺寸内求最大值
                Y[i, j] = X[i: i + p_h, j: j + p_w].max()
            elif mode == 'avg':
                # 池化尺寸内求平均
                Y[i, j] = X[i: i + p_h, j: j + p_w].mean()       
    return Y

X = torch.tensor([[0, 1, 2], [3, 4, 5], [6, 7, 8]])
pool2d(X, (2, 2))

pool2d(X, (2, 2), 'avg')

X = torch.arange(16, dtype=torch.float).view((1, 1, 4, 4))
X

# 只指定池化窗口时，步幅保持一致
pool2d = nn.MaxPool2d(3)
pool2d(X) 

pool2d = nn.MaxPool2d(3, padding=1, stride=2)
pool2d(X)

pool2d = nn.MaxPool2d((2, 4), padding=(1, 2), stride=(2, 3))
pool2d(X)

# torch.cat是将两个张量拼接在一起，dim指定数字为在此维进行拼接
# 原本形状X为(1,1,4,4)，指定拼接为在X.shape[1]维，所以结果为(1,2,4,4)
X = torch.cat((X, X + 1), dim=1)
X

pool2d = nn.MaxPool2d(3, padding=1, stride=2)
pool2d(X)
# (1,2,2,2)

import time
import torch
from torch import nn, optim

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

class LeNet(nn.Module):
    def __init__(self):
        super(LeNet, self).__init__()
        self.conv = nn.Sequential(
            # 输出为6× 24×24
            nn.Conv2d(1, 6, 5), # in_channels, out_channels, kernel_size
            nn.Sigmoid(),
            # 输出为6× 12×12
            nn.MaxPool2d(2, 2), # kernel_size, stride
            # 输出为16× 8×8
            nn.Conv2d(6, 16, 5),
            nn.Sigmoid(),
            # 输出为16× 4×4
            nn.MaxPool2d(2, 2)
        )
        self.fc = nn.Sequential(
            nn.Linear(16*4*4, 120),
            nn.Sigmoid(),
            nn.Linear(120, 84),
            nn.Sigmoid(),
            nn.Linear(84, 10)
        )

    def forward(self, img):
        feature = self.conv(img)
        output = self.fc(feature.view(img.shape[0], -1))
        return output

net = LeNet()
print(net)

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size=batch_size)

# 本函数已保存在d2lzh_pytorch包中方便以后使用。该函数将被逐步改进。
# 此函数在我自己的时候中将其命名添加了后缀变为'evaluate_accuracy_cnn'
# evaluate 评估  accuracy 准确性
def evaluate_accuracy(data_iter, net, 
        device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')):
    acc_sum, n = 0.0, 0
    # 中断梯度追踪
    with torch.no_grad():
        for X, y in data_iter:
            if isinstance(net, torch.nn.Module):
                # eval与tarin涉及到Batch Normalization和Dropout
                # 其中Dropout是3.13中用于减少过拟合的丢弃法
                # Batch Normalization是5.10中的批量归一化
                # 在测试或者评估时丢弃法是不需要的
                # 丢弃法的使用是需要在模型的层中插入nn.Dropout()
                # 需要在其中指定丢弃的概率
                # 位置是只能放在全连接层
                # 某个神经元的激活函数之后
                # 因为它是以指定的丢弃概率随机丢弃上一层的输出元素
                net.eval() # 评估模式, 这会关闭dropout
                # 数据在device中进行计算之后
                # 判断哪些是一样的,结果是一个列表,其中相等的为1,不等的为0
                # float是将其中的类型进行转换
                # sum是将整个列表计算为一个数值
                # cpu表示准确量是在cpu上进行计算.这一步到底是什么考虑不清楚
                # item是将tensor(2.)中取出来只得到  2.0
                # 若是将该变量定义为张量，这一步将变为
                # acc_sum += torch.sum((torch.argmax(net(X), dim=1) == y)).item()
                # 这句话在原文中没有使用y.to(device)，会报错
                acc_sum += (net(X.to(device)).argmax(dim=1) == y.to(device)).float().sum().cpu().item()
                net.train() # 改回训练模式
            # 下面的代码对于本章节是没有意义的
            else: # 自定义的模型, 3.13节之后不会用到, 不考虑GPU
                if('is_training' in net.__code__.co_varnames): # 如果有is_training这个参数
                    # 将is_training设置成False
                    acc_sum += (net(X, is_training=False).argmax(dim=1) == y).float().sum().item() 
                else:
                    acc_sum += (net(X).argmax(dim=1) == y).float().sum().item() 
            n += y.shape[0]
    return acc_sum / n

# 本函数已保存在d2lzh_pytorch包中方便以后使用
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()
    # 训练周期
    for epoch in range(num_epochs):
        # 统计的是数据一共被分成了多少批.每一次train_iter中将所有的样本/批量数就是这个样本分成的批数
    	batch_count = 0
        train_l_sum, train_acc_sum, n, start = 0.0, 0.0, 0, time.time()
        for X, y in train_iter:
            # X形状是(batch_size,c,h,w)
            # y形状是(batch_size)
            X = X.to(device)
            y = y.to(device)
            y_hat = net(X)
            l = loss(y_hat, y)
            optimizer.zero_grad()
            l.backward()
            optimizer.step()
            # 训练损失和是计算每次训练周期的损失值之和
            train_l_sum += l.cpu().item()
            # 训练准确量计算 的是每次训练周期的正确量
            train_acc_sum += (y_hat.argmax(dim=1) == y).sum().cpu().item()
            # y.shape[0]是次此批的数量,由于数据可能不会直接对批量数整除,造成最后一批数量不足
            # 所有n在使用的时候表示的单次训练周期时,所有的样本数量
            n += y.shape[0]
            # 此参数是定义在训练周期循环之外的
            batch_count += 1
        test_acc = evaluate_accuracy(test_iter, net)
        # 根据结果显示的loss，其实使用的是n作为除数了
        #此程序计算的是每批量的损失，但是结果显示的是每元素的损失
        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))

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

import time
import torch
from torch import nn, optim
import torchvision

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

class AlexNet(nn.Module):
    def __init__(self):
        super(AlexNet, self).__init__()
        self.conv = nn.Sequential(
            nn.Conv2d(1, 96, 11, 4), # in_channels, out_channels, kernel_size, stride, padding
            nn.ReLU(),
            nn.MaxPool2d(3, 2), # kernel_size, stride
            # 减小卷积窗口，使用填充为2来使得输入与输出的高和宽一致，且增大输出通道数
            nn.Conv2d(96, 256, 5, 1, 2),
            nn.ReLU(),
            nn.MaxPool2d(3, 2),
            # 连续3个卷积层，且使用更小的卷积窗口。除了最后的卷积层外，进一步增大了输出通道数。
            # 前两个卷积层后不使用池化层来减小输入的高和宽
            nn.Conv2d(256, 384, 3, 1, 1),
            nn.ReLU(),
            nn.Conv2d(384, 384, 3, 1, 1),
            nn.ReLU(),
            nn.Conv2d(384, 256, 3, 1, 1),
            nn.ReLU(),
            nn.MaxPool2d(3, 2)
        )
         # 这里全连接层的输出个数比LeNet中的大数倍。使用丢弃层来缓解过拟合
        self.fc = nn.Sequential(
            nn.Linear(256*5*5, 4096),
            nn.ReLU(),
            nn.Dropout(0.5),
            nn.Linear(4096, 4096),
            nn.ReLU(),
            nn.Dropout(0.5),
            # 输出层。由于这里使用Fashion-MNIST，所以用类别数为10，而非论文中的1000
            nn.Linear(4096, 10),
        )

    def forward(self, img):
        feature = self.conv(img)
        output = self.fc(feature.view(img.shape[0], -1))
        return output

net = AlexNet()
print(net)

# 本函数已保存在d2lzh_pytorch包中方便以后使用
def load_data_fashion_mnist(batch_size, resize=None, root='~/Datasets/FashionMNIST'):
    """Download the fashion mnist dataset and then load into memory."""
    trans = []
    if resize:
        # 重置图像分辨率 输入的要是一个int，将短的数据置为该数值，保持原有的长宽比，计算得出长的数据。若是输入为(h,w),则直接变为该数值。还有一个缺省参数是插值方法，默认是双线性插值
        trans.append(torchvision.transforms.Resize(size=resize))
    # 必须的转为张量，因为输入的数据是PIL Image，若是输入数据为张量则不需要这一步
    trans.append(torchvision.transforms.ToTensor())
    
    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=4)
    test_iter = torch.utils.data.DataLoader(mnist_test, batch_size=batch_size, shuffle=False, num_workers=4)

    return train_iter, test_iter

batch_size = 128
# 如出现“out of memory”的报错信息，可减小batch_size或resize
train_iter, test_iter = load_data_fashion_mnist(batch_size, resize=224)

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

import time
import torch
from torch import nn, optim

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

# 实现多个相同卷积组成的卷积块
def vgg_block(num_convs, in_channels, out_channels):
    blk = []
    # 第一个卷积形状是由输入输出决定的
    for i in range(num_convs):
        if i == 0:
            blk.append(nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1))
        # 之后的保持形状不变，输入输出通道相同
        else:
            blk.append(nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1))
        # 每个卷积之后使用激活函数
        blk.append(nn.ReLU())
    # 每个卷积块之后使用最大池化层，将大小减半
    blk.append(nn.MaxPool2d(kernel_size=2, stride=2)) # 这里会使宽高减半
    # *的作用是将可迭代对象的数据给一下全部“解”出来
    return nn.Sequential(*blk)

conv_arch = ((1, 1, 64), (1, 64, 128), (2, 128, 256), (2, 256, 512), (2, 512, 512))
# 经过5个vgg_block, 宽高会减半5次, 变成 224/32 = 7
fc_features = 512 * 7 * 7 # c * w * h
fc_hidden_units = 4096 # 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(d2l.FlattenLayer(),#这一步是为了将(batch_size,in,h,w)转为(batch_size,in*h*w)
                                 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)
X = torch.rand(1, 1, 224, 224)

# named_children获取一级子模块及其名字(named_modules会返回所有子模块,包括子模块的子模块)
for name, blk in net.named_children(): 
    X = blk(X)
    print(name, 'output shape: ', X.shape)

ratio = 8
small_conv_arch = [(1, 1, 64//ratio), (1, 64//ratio, 128//ratio), (2, 128//ratio, 256//ratio), 
                   (2, 256//ratio, 512//ratio), (2, 512//ratio, 512//ratio)]
net = vgg(small_conv_arch, fc_features // ratio, fc_hidden_units // ratio)
print(net)

batch_size = 64
# 如出现“out of memory”的报错信息，可减小batch_size或resize
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size, resize=224)

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

import time
import torch
from torch import nn, optim

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

# 一个卷积和两个1×1卷积组成
def nin_block(in_channels, out_channels, kernel_size, stride, padding):
    blk = nn.Sequential(
        nn.Conv2d(in_channels, out_channels, kernel_size, stride, padding),
        nn.ReLU(),
        nn.Conv2d(out_channels, out_channels, kernel_size=1),
        nn.ReLU(),
        nn.Conv2d(out_channels, out_channels, kernel_size=1),
        nn.ReLU())
    return blk

# 已保存在d2lzh_pytorch
# functional是可直接调用的函数，nn是封装好的类
import torch.nn.functional as F
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:])

net = nn.Sequential(
    nin_block(1, 96, kernel_size=11, stride=4, padding=0),
    nn.MaxPool2d(kernel_size=3, stride=2),
    nin_block(96, 256, kernel_size=5, stride=1, padding=2),
    nn.MaxPool2d(kernel_size=3, stride=2),
    nin_block(256, 384, kernel_size=3, stride=1, padding=1),
    nn.MaxPool2d(kernel_size=3, stride=2), 
    nn.Dropout(0.5),
    # 标签类别数是10
    nin_block(384, 10, kernel_size=3, stride=1, padding=1),
    GlobalAvgPool2d(), 
    # 将四维的输出转成二维的输出，其形状为(批量大小, 10)
    d2l.FlattenLayer())

X = torch.rand(1, 1, 224, 224)
for name, blk in net.named_children(): 
    X = blk(X)
    print(name, 'output shape: ', X.shape)

batch_size = 128
# 如出现“out of memory”的报错信息，可减小batch_size或resize
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size, resize=224)

lr, num_epochs = 0.002, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

import time
import torch
from torch import nn, optim
import torch.nn.functional as F

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

class Inception(nn.Module):
    # c1 - c4为每条线路里的层的输出通道数
    def __init__(self, in_c, c1, c2, c3, c4):
        super(Inception, self).__init__()
        # 线路1，单1 x 1卷积层
        self.p1_1 = nn.Conv2d(in_c, c1, kernel_size=1)
        # 线路2，1 x 1卷积层后接3 x 3卷积层
        self.p2_1 = nn.Conv2d(in_c, c2[0], kernel_size=1)
        self.p2_2 = nn.Conv2d(c2[0], c2[1], kernel_size=3, padding=1)
        # 线路3，1 x 1卷积层后接5 x 5卷积层
        self.p3_1 = nn.Conv2d(in_c, c3[0], kernel_size=1)
        self.p3_2 = nn.Conv2d(c3[0], c3[1], kernel_size=5, padding=2)
        # 线路4，3 x 3最大池化层后接1 x 1卷积层
        self.p4_1 = nn.MaxPool2d(kernel_size=3, stride=1, padding=1)
        self.p4_2 = nn.Conv2d(in_c, c4, kernel_size=1)

    def forward(self, x):
        # 在每个卷积块都是用了Relu激活函数
        p1 = F.relu(self.p1_1(x))
        p2 = F.relu(self.p2_2(F.relu(self.p2_1(x))))
        p3 = F.relu(self.p3_2(F.relu(self.p3_1(x))))
        p4 = F.relu(self.p4_2(self.p4_1(x)))
        return torch.cat((p1, p2, p3, p4), dim=1)  # 在通道维上连结输出

b1 = nn.Sequential(nn.Conv2d(1, 64, kernel_size=7, stride=2, padding=3),
                   nn.ReLU(),
                   nn.MaxPool2d(kernel_size=3, stride=2, padding=1))

b2 = nn.Sequential(nn.Conv2d(64, 64, kernel_size=1),
                   nn.Conv2d(64, 192, kernel_size=3, padding=1),
                   nn.MaxPool2d(kernel_size=3, stride=2, padding=1))

b3 = nn.Sequential(Inception(192, 64, (96, 128), (16, 32), 32),
                   Inception(256, 128, (128, 192), (32, 96), 64),
                   nn.MaxPool2d(kernel_size=3, stride=2, padding=1))

b4 = nn.Sequential(Inception(480, 192, (96, 208), (16, 48), 64),
                   Inception(512, 160, (112, 224), (24, 64), 64),
                   Inception(512, 128, (128, 256), (24, 64), 64),
                   Inception(512, 112, (144, 288), (32, 64), 64),
                   Inception(528, 256, (160, 320), (32, 128), 128),
                   nn.MaxPool2d(kernel_size=3, stride=2, padding=1))

b5 = nn.Sequential(Inception(832, 256, (160, 320), (32, 128), 128),
                   Inception(832, 384, (192, 384), (48, 128), 128),
                   d2l.GlobalAvgPool2d())

net = nn.Sequential(b1, b2, b3, b4, b5, 
                    d2l.FlattenLayer(), nn.Linear(1024, 10))

net = nn.Sequential(b1, b2, b3, b4, b5, d2l.FlattenLayer(), nn.Linear(1024, 10))
# （批量数，通道数，高，宽）
X = torch.rand(1, 1, 96, 96)
# children()返回当前模型 子模块的迭代器
for blk in net.children(): 
    X = blk(X)
    print('output shape: ', X.shape)

batch_size = 128
# 如出现“out of memory”的报错信息，可减小batch_size或resize
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size, resize=96)

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

import time
import torch
from torch import nn, optim
import torch.nn.functional as F

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

def batch_norm(is_training, X, gamma, beta, moving_mean, moving_var, eps, momentum):
    # 判断当前模式是训练模式还是预测模式
    if not is_training:
        # 如果是在预测模式下，直接使用传入的移动平均所得的均值和方差
        X_hat = (X - moving_mean) / torch.sqrt(moving_var + eps)
    else:
        assert len(X.shape) in (2, 4)
        if len(X.shape) == 2:
            # 使用全连接层的情况，计算特征维上的均值和方差
            mean = X.mean(dim=0)
            var = ((X - mean) ** 2).mean(dim=0)
        else:
            # 使用二维卷积层的情况，计算通道维上（axis=1）的均值和方差。这里我们需要保持
            # X的形状以便后面可以做广播运算
            mean = X.mean(dim=0, keepdim=True).mean(dim=2, keepdim=True).mean(dim=3, keepdim=True)
            var = ((X - mean) ** 2).mean(dim=0, keepdim=True).mean(dim=2, keepdim=True).mean(dim=3, keepdim=True)
        # 训练模式下用当前的均值和方差做标准化
        X_hat = (X - mean) / torch.sqrt(var + eps)
        # 更新移动平均的均值和方差
        moving_mean = momentum * moving_mean + (1.0 - momentum) * mean
        moving_var = momentum * moving_var + (1.0 - momentum) * var
    Y = gamma * X_hat + beta  # 拉伸和偏移
    return Y, moving_mean, moving_var

class BatchNorm(nn.Module):
    def __init__(self, num_features, num_dims):
        super(BatchNorm, self).__init__()
        if num_dims == 2:
            shape = (1, num_features)
        else:
            shape = (1, num_features, 1, 1)
        # 参与求梯度和迭代的拉伸和偏移参数，分别初始化成0和1
        self.gamma = nn.Parameter(torch.ones(shape))
        self.beta = nn.Parameter(torch.zeros(shape))
        # 不参与求梯度和迭代的变量，全在内存上初始化成0
        self.moving_mean = torch.zeros(shape)
        self.moving_var = torch.zeros(shape)

    def forward(self, X):
        # 如果X不在内存上，将moving_mean和moving_var复制到X所在显存上
        if self.moving_mean.device != X.device:
            self.moving_mean = self.moving_mean.to(X.device)
            self.moving_var = self.moving_var.to(X.device)
        # 保存更新过的moving_mean和moving_var, Module实例的traning属性默认为true, 调用.eval()后设成false
        Y, self.moving_mean, self.moving_var = batch_norm(self.training, 
            X, self.gamma, self.beta, self.moving_mean,
            self.moving_var, eps=1e-5, momentum=0.9)
        return Y

net = nn.Sequential(
            nn.Conv2d(1, 6, 5), # in_channels, out_channels, kernel_size
            BatchNorm(6, num_dims=4),
            nn.Sigmoid(),
            nn.MaxPool2d(2, 2), # kernel_size, stride
            nn.Conv2d(6, 16, 5),
            BatchNorm(16, num_dims=4),
            nn.Sigmoid(),
            nn.MaxPool2d(2, 2),
            d2l.FlattenLayer(),
            nn.Linear(16*4*4, 120),
            BatchNorm(120, num_dims=2),
            nn.Sigmoid(),
            nn.Linear(120, 84),
            BatchNorm(84, num_dims=2),
            nn.Sigmoid(),
            nn.Linear(84, 10)
        )

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size=batch_size)

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

net[1].gamma.view((-1,)), net[1].beta.view((-1,))

net = nn.Sequential(
            nn.Conv2d(1, 6, 5), # in_channels, out_channels, kernel_size
            nn.BatchNorm2d(6),
            nn.Sigmoid(),
            nn.MaxPool2d(2, 2), # kernel_size, stride
            nn.Conv2d(6, 16, 5),
            nn.BatchNorm2d(16),
            nn.Sigmoid(),
            nn.MaxPool2d(2, 2),
            d2l.FlattenLayer(),
            nn.Linear(16*4*4, 120),
            nn.BatchNorm1d(120),
            nn.Sigmoid(),
            nn.Linear(120, 84),
            nn.BatchNorm1d(84),
            nn.Sigmoid(),
            nn.Linear(84, 10)
        )

batch_size = 256
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size=batch_size)

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)

import time
import torch
from torch import nn, optim
import torch.nn.functional as F

import sys
sys.path.append("..") 
import d2lzh_pytorch as d2l
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

class Residual(nn.Module):  # 本类已保存在d2lzh_pytorch包中方便以后使用
    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)

blk = Residual(3, 3)
# 批量数，输入通道，h,w
X = torch.rand((4, 3, 6, 6))
blk(X).shape # torch.Size([4, 3, 6, 6])

blk = Residual(3, 6, use_1x1conv=True, stride=2)
blk(X).shape # torch.Size([4, 6, 3, 3])

net = nn.Sequential(
        nn.Conv2d(1, 64, kernel_size=7, stride=2, padding=3),
        nn.BatchNorm2d(64), 
        nn.ReLU(),
        nn.MaxPool2d(kernel_size=3, stride=2, padding=1))

def resnet_block(in_channels, out_channels, num_residuals, first_block=False):
    if first_block:
        #assert 断言可以在条件不满足程序运行的情况下直接返回错误，而不必等待程序运行后出现崩溃的情况
        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.add_module("resnet_block1", resnet_block(64, 64, 2, first_block=True))
net.add_module("resnet_block2", resnet_block(64, 128, 2))
net.add_module("resnet_block3", resnet_block(128, 256, 2))
net.add_module("resnet_block4", resnet_block(256, 512, 2))

net.add_module("global_avg_pool", d2l.GlobalAvgPool2d()) # GlobalAvgPool2d的输出: (Batch, 512, 1, 1)
net.add_module("fc", nn.Sequential(d2l.FlattenLayer(), nn.Linear(512, 10))) 

X = torch.rand((1, 1, 224, 224))
for name, layer in net.named_children():
    X = layer(X)
    print(name, ' output shape:\t', X.shape)

batch_size = 256
# 如出现“out of memory”的报错信息，可减小batch_size或resize
train_iter, test_iter = d2l.load_data_fashion_mnist(batch_size, resize=96)

lr, num_epochs = 0.001, 5
optimizer = torch.optim.Adam(net.parameters(), lr=lr)
d2l.train_ch5(net, train_iter, test_iter, batch_size, optimizer, device, num_epochs)