莫烦的网址这里
关于卷积神经网络的介绍:
自行搜索或者参考博客1,在莫烦中也有相应的视频
我们接下来直接手动实现一个简单的数字识别的CNN网络:
1.首先定义数据集
通过pytorch的torchvision.datasets网址
dset.MNIST(root, train=True, transform=None, target_transform=None, download=False)
参数说明: - root : processed/training.pt
和 processed/test.pt
的主目录 - train : True
= 训练集, False
= 测试集 - download : True
= 从互联网上下载数据集,并把数据集放在root
目录下. 如果数据集之前下载过,将处理过的数据(minist.py中有相关函数)放在processed
文件夹下, - transform
: 一个函数,原始图片作为输入,返回一个转换后的图片。
target_transform
- 一个函数,输入为target
,输出对其的转换。例子,输入的是图片标注的string
,输出为word
的索引。
1.我们建立test数据集:
train_data = torchvision.datasets.MNIST(
root='./mnist/',
train=True, # this is training data
transform=torchvision.transforms.ToTensor(), # 黑白图片只有一个层,变成值0-1,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, #是否下载?
#download=False #已经下载
)
2.建立测试数据集,进行评价的时候用
test_data = torchvision.datasets.MNIST(root='./mnist/', train=False) #FALSE说明是testdata
test_x = torch.unsqueeze(test_data.test_data, dim=1).type(torch.FloatTensor)[:2000]/255. # shape from (2000, 28, 28) to (2000, 1, 28, 28), value in range(0,1)
test_y = test_data.test_labels[:2000]
3.然后定义包装数据和tensor的抽象数据集
train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)
该函数参考网址,数据加载器。组合数据集和采样器,并在数据集上提供单进程或多进程迭代器。
参数:
- dataset (Dataset) – 加载数据的数据集。
- batch_size (int, optional) – 每个batch加载多少个样本(默认: 1)。
- shuffle (bool, optional) – 设置为True时会在每个epoch重新打乱数据(默认: False)
2.创建CNN神经网络
首先CNN网络的model是继承了torch.nn.Moudle,
class CNN(nn.Module):
def __init__(self):
super(CNN, self).__init__()
self.conv1 = nn.Sequential( # 卷积层,input shape (1, 28, 28)
nn.Conv2d( #卷积层收集器,收集信息
in_channels=1, # input height,黑白图片
out_channels=16, # n_filters,
kernel_size=5, # filter size
stride=1, # filter movement/step,每隔多少步跳跃
padding=2, # 如果不能整除channelssize,添加一层0数据,if want same width and length of this image after Conv2d, padding=(kernel_size-1)/2 if stride=1
), # output shape (16, 28, 28)
nn.ReLU(), # activation
nn.MaxPool2d(kernel_size=2), # 筛选choose max value in 2x2 area, output shape (16, 14, 14)
)
self.conv2 = nn.Sequential( # input shape (16, 14, 14)
nn.Conv2d(16, 32, 5, 1, 2), # 相当于上面的变量,output shape (32, 14, 14)
nn.ReLU(), # activation
nn.MaxPool2d(2), # Max选取最大值,Avgpool2d选平均,utput shape (32, 7, 7)
)
self.out = nn.Linear(32 * 7 * 7, 10) # fully connected layer, output 10 classes (0-9十个分类)
def forward(self, x):
x = self.conv1(x)
x = self.conv2(x) #batch(32,7,7)
x = x.view(x.size(0), -1) #展平 flatten the output of conv2 to (batch_size, 32 * 7 * 7)
output = self.out(x)
return output, x # return x for visualization
而我们这个,我们建立了两个时序容器:每个时序容器中包含了卷积层和激励函数以及池化层:
其中的Conv2d是一个卷积层:class torch.nn.Conv2d(in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=True)二维卷积层, 输入的尺度是(N, C_in,H,W),输出尺度(N,C_out,H_out,W_out)的计算方式,其具体参数可以查看手册手册#1 ,这里仅仅使用了以下参数:
- in_channels(
int
) – 输入信号的通道 - out_channels(
int
) – 卷积产生的通道 - kerner_size(
int
ortuple
) - 卷积核的尺寸 - stride(
int
ortuple
,optional
) - 卷积步长 - padding (
int
ortuple
,optional
)- 输入的每一条边补充0的层数
其中MaxPool2d 是池化层,class torch.nn.MaxPool2d(kernel_size, stride=None, padding=0, dilation=1, return_indices=False, ceil_mode=False),对于输入信号的输入通道,提供2维最大池化(max pooling
)操作 这里只改了参数:kernel_size(int
or tuple
) - max pooling的窗口大小
3.训练神经网络
其原理依旧是类似之前的,不详述了。
至于其显示测试的细节部分的代码,emmmm,与CNN关联不大,自行学习
所有代码如下:
"""
Dependencies:
torch: 0.4
torchvision
matplotlib
"""
# library
# standard library
import os
# third-party library
import torch
import torch.nn as nn
import torch.utils.data as Data
import torchvision
import matplotlib.pyplot as plt
# torch.manual_seed(1) # reproducible
# Hyper Parameters
EPOCH = 1 # train the training data n times, to save time, we just train 1 epoch
BATCH_SIZE = 50
LR = 0.001 # learning rate
DOWNLOAD_MNIST = False
# Mnist digits dataset
if not(os.path.exists('./mnist/')) or not os.listdir('./mnist/'):
# not mnist dir or mnist is empyt dir
DOWNLOAD_MNIST = True
train_data = torchvision.datasets.MNIST(
root='./mnist/',
train=True, # this is training data
transform=torchvision.transforms.ToTensor(), # 黑白图片只有一个层,变成值0-1,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, #是否下载?
download=False #已经下载
)
# plot one example
# 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[0])
# plt.show()
# Data Loader for easy mini-batch return in training, the image batch shape will be (50, 1, 28, 28)
train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)
# pick 2000 samples to speed up testing
test_data = torchvision.datasets.MNIST(root='./mnist/', train=False) #FALSE说明是testdata
test_x = torch.unsqueeze(test_data.test_data, dim=1).type(torch.FloatTensor)[:2000]/255. # shape from (2000, 28, 28) to (2000, 1, 28, 28), value in range(0,1)
test_y = test_data.test_labels[:2000]
class CNN(nn.Module):
def __init__(self):
super(CNN, self).__init__()
self.conv1 = nn.Sequential( # 卷积层,input shape (1, 28, 28)
nn.Conv2d( #卷积层收集器,收集信息
in_channels=1, # input height,黑白图片
out_channels=16, # n_filters,
kernel_size=5, # filter size
stride=1, # filter movement/step,每隔多少步跳跃
padding=2, # 如果不能整除channelssize,添加一层0数据,if want same width and length of this image after Conv2d, padding=(kernel_size-1)/2 if stride=1
), # output shape (16, 28, 28)
nn.ReLU(), # activation
nn.MaxPool2d(kernel_size=2), # 筛选choose max value in 2x2 area, output shape (16, 14, 14)
)
self.conv2 = nn.Sequential( # input shape (16, 14, 14)
nn.Conv2d(16, 32, 5, 1, 2), # 相当于上面的变量,output shape (32, 14, 14)
nn.ReLU(), # activation
nn.MaxPool2d(2), # Max选取最大值,Avgpool2d选平均,utput shape (32, 7, 7)
)
self.out = nn.Linear(32 * 7 * 7, 10) # fully connected layer, output 10 classes (0-9十个分类)
def forward(self, x):
x = self.conv1(x)
x = self.conv2(x) #batch(32,7,7)
x = x.view(x.size(0), -1) #展平 flatten the output of conv2 to (batch_size, 32 * 7 * 7)
output = self.out(x)
return output, x # return x for visualization
cnn = CNN()
print(cnn) # net architecture
optimizer = torch.optim.Adam(cnn.parameters(), lr=LR) # optimize all cnn parameters
loss_func = nn.CrossEntropyLoss() # the target label is not one-hotted
# following function (plot_with_labels) is for visualization, can be ignored if not interested
from matplotlib import cm
try: from sklearn.manifold import TSNE; HAS_SK = True
except: HAS_SK = False; print('Please install sklearn for layer visualization')
def plot_with_labels(lowDWeights, labels):
plt.cla()
X, Y = lowDWeights[:, 0], lowDWeights[:, 1]
for x, y, s in zip(X, Y, labels):
c = cm.rainbow(int(255 * s / 9)); plt.text(x, y, s, backgroundcolor=c, fontsize=9)
plt.xlim(X.min(), X.max()); plt.ylim(Y.min(), Y.max()); plt.title('Visualize last layer'); plt.show(); plt.pause(0.01)
plt.ion()
# training and testing
for epoch in range(EPOCH):
for step, (b_x, b_y) in enumerate(train_loader): # gives batch data, normalize x when iterate train_loader
output = cnn(b_x)[0] # cnn output
loss = loss_func(output, b_y) # cross entropy loss
optimizer.zero_grad() # clear gradients for this training step
loss.backward() # backpropagation, compute gradients
optimizer.step() # apply gradients
if step % 50 == 0:
test_output, last_layer = cnn(test_x)
pred_y = torch.max(test_output, 1)[1].data.numpy()
accuracy = float((pred_y == test_y.data.numpy()).astype(int).sum()) / float(test_y.size(0))
print('Epoch: ', epoch, '| train loss: %.4f' % loss.data.numpy(), '| test accuracy: %.2f' % accuracy)
if HAS_SK:
# Visualization of trained flatten layer (T-SNE)
tsne = TSNE(perplexity=30, n_components=2, init='pca', n_iter=5000)
plot_only = 500
low_dim_embs = tsne.fit_transform(last_layer.data.numpy()[:plot_only, :])
labels = test_y.numpy()[:plot_only]
plot_with_labels(low_dim_embs, labels)
plt.ioff()
# print 10 predictions from test data
test_output, _ = cnn(test_x[:10])
pred_y = torch.max(test_output, 1)[1].data.numpy()
print(pred_y, 'prediction number')
print(test_y[:10].numpy(), 'real number')