pytorch实现自编码

自编码网络是非监督学习领域的一种，可以自动从无标注的数据中学习特征，是一种以重构输入信息为目标的神经网络，它可以给出比原始数据更好的特征描述，具有较强的特征学习能力，在深度学习中常用自编码网络生成的特征来取代原始数据，已取得更好效果。

换句话说，自编码网络的作用相当于PCA并且能获得比PCA更好的效果

这里我们把手写数据集里面的图片特征先压缩，根据压缩出来的特征，对数据进行分类，即无监督学习（无监督学习：不需要标签，只需要用到train_x，不需要用到train_y）

import torch.nn as nn
import torch.utils.data as Data
import torchvision
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import cm
import numpy as np

#超参数
EPOCH = 10
BATCH_SIZE = 64
LR = 0.005         # learning rate
DOWNLOAD_MNIST = False
N_TEST_IMG = 5

# Mnist digits dataset
train_data = torchvision.datasets.MNIST(
    root=r'D:\python\minist',
    train=True,                                     # this is training data
    transform=torchvision.transforms.ToTensor(),    # Converts a PIL.Image or numpy.ndarray to
                                                    # torch.FloatTensor of shape (C x H x W) and normalize in the range [0.0, 1.0]
    download=DOWNLOAD_MNIST,                        # 如果已经有了就填False，没有就填Ture
)

# 查看一个数据
print(train_data.train_data.size())     # (60000, 28, 28)
print(train_data.train_labels.size())   # (60000)
plt.imshow(train_data.train_data[0].numpy(), cmap='gray')
plt.title('%i' % train_data.train_labels[2])
plt.show()

得到其中一个数据的显示

#将数据小批量训练，图像数据变成 (50, 1, 28, 28)
train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)


#构造神经网络
class AutoEncoder(nn.Module):
    def __init__(self):
        super(AutoEncoder, self).__init__()

        self.encoder = nn.Sequential(
            nn.Linear(28*28, 128),
            nn.Tanh(), #激活函数
            nn.Linear(128, 64),
            nn.Tanh(),
            nn.Linear(64, 12),
            nn.Tanh(),
            nn.Linear(12, 3),   # compress to 3 features which can be visualized in plt
        )
        
        #定义解压器
        self.decoder = nn.Sequential(
            nn.Linear(3, 12),
            nn.Tanh(),
            nn.Linear(12, 64),
            nn.Tanh(),
            nn.Linear(64, 128),
            nn.Tanh(),
            nn.Linear(128, 28*28),
            nn.Sigmoid(),       # 把输出值控制到0-1的范围
        )

    def forward(self, x):
        encoded = self.encoder(x)
        decoded = self.decoder(encoded)
        return encoded, decoded

#自编码网络实例化
autoencoder = AutoEncoder()


#定义优化器和损失函数
optimizer = torch.optim.Adam(autoencoder.parameters(), lr=LR)
loss_func = nn.MSELoss()

首先画一个2*5的图

#显示原始图片
f, a = plt.subplots(2, N_TEST_IMG, figsize=(5, 2))
plt.ion()   

第一行显示测试数据

#显示第一行数据
view_data = train_data.train_data[:N_TEST_IMG].view(-1, 28*28).type(torch.FloatTensor)/255.
for i in range(N_TEST_IMG):
    a[0][i].imshow(np.reshape(view_data.data.numpy()[i], (28, 28)), cmap='gray'); a[0][i].set_xticks(()); a[0][i].set_yticks(())

#训练过程
for epoch in range(EPOCH):
    for step, (x, b_label) in enumerate(train_loader):
        b_x = x.view(-1, 28*28)   # batch x, 把x reshape成(batch, 28*28) 
        b_y = x.view(-1, 28*28)   # batch y, 在数据上和b_x一样

        encoded, decoded = autoencoder(b_x)

        loss = loss_func(decoded, b_y)      # 对比原图片和解压出来的图片的误差
        optimizer.zero_grad()               # clear gradients for this training step
        loss.backward()                     # backpropagation, compute gradients
        optimizer.step()                    # apply gradients

        if step % 100 == 0: #每一百步画一次图，并且画出解压的图片
            print('Epoch: ', epoch, '| train loss: %.4f' % loss.data.numpy())

            # 在第二行画解压出来的图片
            _, decoded_data = autoencoder(view_data)
            for i in range(N_TEST_IMG):
                a[1][i].clear()
                a[1][i].imshow(np.reshape(decoded_data.data.numpy()[i], (28, 28)), cmap='gray')
                a[1][i].set_xticks(()); a[1][i].set_yticks(())
            plt.draw(); plt.pause(0.05)

plt.ioff()
plt.show()

画出三维分类图：

# visualize in 3D plot
view_data = train_data.train_data[:200].view(-1, 28*28).type(torch.FloatTensor)/255.
encoded_data, _ = autoencoder(view_data)
fig = plt.figure(2); ax = Axes3D(fig)
X, Y, Z = encoded_data.data[:, 0].numpy(), encoded_data.data[:, 1].numpy(), encoded_data.data[:, 2].numpy()
#这里的X，Y，Z表示编码之后的前三个属性
values = train_data.train_labels[:200].numpy()
for x, y, z, s in zip(X, Y, Z, values):
    c = cm.rainbow(int(255*s/9)); ax.text(x, y, z, s, backgroundcolor=c)
ax.set_xlim(X.min(), X.max()); ax.set_ylim(Y.min(), Y.max()); ax.set_zlim(Z.min(), Z.max())
plt.show()

完整：

import torch
import torch.nn as nn
import torch.utils.data as Data
import torchvision
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import cm
import numpy as np

#超参数
EPOCH = 10
BATCH_SIZE = 64
LR = 0.005         # learning rate
DOWNLOAD_MNIST = False
N_TEST_IMG = 5 #测试数据是五个

# Mnist digits dataset
train_data = torchvision.datasets.MNIST(
    root=r'D:\python\minist',
    train=True,                                     # this is training data
    transform=torchvision.transforms.ToTensor(),    # Converts a PIL.Image or numpy.ndarray to
                                                    # torch.FloatTensor of shape (C x H x W) and normalize in the range [0.0, 1.0]
    download=DOWNLOAD_MNIST,                        # 如果已经有了就填False，没有就填Ture
)

## 查看一个数据
#print(train_data.train_data.size())     # (60000, 28, 28)
#print(train_data.train_labels.size())   # (60000)
#plt.imshow(train_data.train_data[0].numpy(), cmap='gray')
#plt.title('%i' % train_data.train_labels[2])
#plt.show()


#将数据小批量训练，图像数据变成 (50, 1, 28, 28)
train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)


#构造神经网络
class AutoEncoder(nn.Module):
    def __init__(self):
        super(AutoEncoder, self).__init__()

        self.encoder = nn.Sequential(
            nn.Linear(28*28, 128),
            nn.Tanh(), #激活函数
            nn.Linear(128, 64),
            nn.Tanh(),
            nn.Linear(64, 12),
            nn.Tanh(),
            nn.Linear(12, 3),   # compress to 3 features which can be visualized in plt
        )
        
        #定义解压器
        self.decoder = nn.Sequential(
            nn.Linear(3, 12),
            nn.Tanh(),
            nn.Linear(12, 64),
            nn.Tanh(),
            nn.Linear(64, 128),
            nn.Tanh(),
            nn.Linear(128, 28*28),
            nn.Sigmoid(),       # 把输出值控制到0-1的范围
        )

    def forward(self, x):
        encoded = self.encoder(x)
        decoded = self.decoder(encoded)
        return encoded, decoded

#自编码网络实例化
autoencoder = AutoEncoder()


#定义优化器和损失函数
optimizer = torch.optim.Adam(autoencoder.parameters(), lr=LR)
loss_func = nn.MSELoss()


#显示原始图片
f, a = plt.subplots(2, N_TEST_IMG, figsize=(5, 2))
plt.ion()   
#显示第一行数据
view_data = train_data.train_data[:N_TEST_IMG].view(-1, 28*28).type(torch.FloatTensor)/255 #训练集的前五个数字
for i in range(N_TEST_IMG):
    a[0][i].imshow(np.reshape(view_data.data.numpy()[i], (28, 28)), cmap='gray'); a[0][i].set_xticks(()); a[0][i].set_yticks(())
    


#训练过程
for epoch in range(EPOCH):
    for step, (x, b_label) in enumerate(train_loader):
        b_x = x.view(-1, 28*28)   # batch x, 把x reshape成(batch, 28*28) 
        b_y = x.view(-1, 28*28)   # batch y, 在数据上和b_x一样

        encoded, decoded = autoencoder(b_x)

        loss = loss_func(decoded, b_y)      # 对比原图片和解压出来的图片的误差
        optimizer.zero_grad()               # clear gradients for this training step
        loss.backward()                     # backpropagation, compute gradients
        optimizer.step()                    # apply gradients

        if step % 100 == 0: #每一百步画一次图，并且画出解压的图片
            print('Epoch: ', epoch, '| train loss: %.4f' % loss.data.numpy())

            # 在第二行画解压出来的图片
            _, decoded_data = autoencoder(view_data)
            for i in range(N_TEST_IMG):
                a[1][i].clear()
                a[1][i].imshow(np.reshape(decoded_data.data.numpy()[i], (28, 28)), cmap='gray')
                a[1][i].set_xticks(()); a[1][i].set_yticks(())
            plt.draw(); plt.pause(0.05)

plt.ioff()
plt.show()

# visualize in 3D plot
view_data = train_data.train_data[:200].view(-1, 28*28).type(torch.FloatTensor)/255.
encoded_data, _ = autoencoder(view_data)
fig = plt.figure(2); ax = Axes3D(fig)
X, Y, Z = encoded_data.data[:, 0].numpy(), encoded_data.data[:, 1].numpy(), encoded_data.data[:, 2].numpy()
#这里的X，Y，Z表示编码之后的前三个属性
values = train_data.train_labels[:200].numpy()
for x, y, z, s in zip(X, Y, Z, values):
    c = cm.rainbow(int(255*s/9)); ax.text(x, y, z, s, backgroundcolor=c)
ax.set_xlim(X.min(), X.max()); ax.set_ylim(Y.min(), Y.max()); ax.set_zlim(Z.min(), Z.max())
plt.show()