NNDL 作业6：基于CNN的XO识别

一、实现卷积-池化-激活

1.Nmupy版本：手工实现 卷积-池化-激活

import numpy as np
 
x = np.array([[-1, -1, -1, -1, -1, -1, -1, -1, -1],
              [-1, 1, -1, -1, -1, -1, -1, 1, -1],
              [-1, -1, 1, -1, -1, -1, 1, -1, -1],
              [-1, -1, -1, 1, -1, 1, -1, -1, -1],
              [-1, -1, -1, -1, 1, -1, -1, -1, -1],
              [-1, -1, -1, 1, -1, 1, -1, -1, -1],
              [-1, -1, 1, -1, -1, -1, 1, -1, -1],
              [-1, 1, -1, -1, -1, -1, -1, 1, -1],
              [-1, -1, -1, -1, -1, -1, -1, -1, -1]])
print("x=\n", x)
# 初始化 三个 卷积核
Kernel = [[0 for i in range(0, 3)] for j in range(0, 3)]
Kernel[0] = np.array([[1, -1, -1],
                      [-1, 1, -1],
                      [-1, -1, 1]])
Kernel[1] = np.array([[1, -1, 1],
                      [-1, 1, -1],
                      [1, -1, 1]])
Kernel[2] = np.array([[-1, -1, 1],
                      [-1, 1, -1],
                      [1, -1, -1]])
 
# --------------- 卷积  ---------------
stride = 1  # 步长
feature_map_h = 7  # 特征图的高
feature_map_w = 7  # 特征图的宽
feature_map = [0 for i in range(0, 3)]  # 初始化3个特征图
for i in range(0, 3):
    feature_map[i] = np.zeros((feature_map_h, feature_map_w))  # 初始化特征图
for h in range(feature_map_h):  # 向下滑动，得到卷积后的固定行
    for w in range(feature_map_w):  # 向右滑动，得到卷积后的固定行的列
        v_start = h * stride  # 滑动窗口的起始行（高）
        v_end = v_start + 3  # 滑动窗口的结束行（高）
        h_start = w * stride  # 滑动窗口的起始列（宽）
        h_end = h_start + 3  # 滑动窗口的结束列（宽）
        window = x[v_start:v_end, h_start:h_end]  # 从图切出一个滑动窗口
        for i in range(0, 3):
            feature_map[i][h, w] = np.divide(np.sum(np.multiply(window, Kernel[i][:, :])), 9)
print("feature_map:\n", np.around(feature_map, decimals=2))
 
# --------------- 池化  ---------------
pooling_stride = 2  # 步长
pooling_h = 4  # 特征图的高
pooling_w = 4  # 特征图的宽
feature_map_pad_0 = [[0 for i in range(0, 8)] for j in range(0, 8)]
for i in range(0, 3):  # 特征图 补 0 ，行 列 都要加 1 (因为上一层是奇数，池化窗口用的偶数)
    feature_map_pad_0[i] = np.pad(feature_map[i], ((0, 1), (0, 1)), 'constant', constant_values=(0, 0))
# print("feature_map_pad_0 0:\n", np.around(feature_map_pad_0[0], decimals=2))
 
pooling = [0 for i in range(0, 3)]
for i in range(0, 3):
    pooling[i] = np.zeros((pooling_h, pooling_w))  # 初始化特征图
for h in range(pooling_h):  # 向下滑动，得到卷积后的固定行
    for w in range(pooling_w):  # 向右滑动，得到卷积后的固定行的列
        v_start = h * pooling_stride  # 滑动窗口的起始行（高）
        v_end = v_start + 2  # 滑动窗口的结束行（高）
        h_start = w * pooling_stride  # 滑动窗口的起始列（宽）
        h_end = h_start + 2  # 滑动窗口的结束列（宽）
        for i in range(0, 3):
            pooling[i][h, w] = np.max(feature_map_pad_0[i][v_start:v_end, h_start:h_end])
print("pooling:\n", np.around(pooling[0], decimals=2))
print("pooling:\n", np.around(pooling[1], decimals=2))
print("pooling:\n", np.around(pooling[2], decimals=2))
 
 
# --------------- 激活  ---------------
def relu(x):
    return (abs(x) + x) / 2
 
 
relu_map_h = 7  # 特征图的高
relu_map_w = 7  # 特征图的宽
relu_map = [0 for i in range(0, 3)]  # 初始化3个特征图
for i in range(0, 3):
    relu_map[i] = np.zeros((relu_map_h, relu_map_w))  # 初始化特征图
 
for i in range(0, 3):
    relu_map[i] = relu(feature_map[i])
 
print("relu map :\n",np.around(relu_map[0], decimals=2))
print("relu map :\n",np.around(relu_map[1], decimals=2))
print("relu map :\n",np.around(relu_map[2], decimals=2))

实验结果：

x=
 [[-1 -1 -1 -1 -1 -1 -1 -1 -1]
 [-1  1 -1 -1 -1 -1 -1  1 -1]
 [-1 -1  1 -1 -1 -1  1 -1 -1]
 [-1 -1 -1  1 -1  1 -1 -1 -1]
 [-1 -1 -1 -1  1 -1 -1 -1 -1]
 [-1 -1 -1  1 -1  1 -1 -1 -1]
 [-1 -1  1 -1 -1 -1  1 -1 -1]
 [-1  1 -1 -1 -1 -1 -1  1 -1]
 [-1 -1 -1 -1 -1 -1 -1 -1 -1]]
feature_map:
 [[[ 0.78 -0.11  0.11  0.33  0.56 -0.11  0.33]
  [-0.11  1.   -0.11  0.33 -0.11  0.11 -0.11]
  [ 0.11 -0.11  1.   -0.33  0.11 -0.11  0.56]
  [ 0.33  0.33 -0.33  0.56 -0.33  0.33  0.33]
  [ 0.56 -0.11  0.11 -0.33  1.   -0.11  0.11]
  [-0.11  0.11 -0.11  0.33 -0.11  1.   -0.11]
  [ 0.33 -0.11  0.56  0.33  0.11 -0.11  0.78]]

 [[ 0.33 -0.56  0.11 -0.11  0.11 -0.56  0.33]
  [-0.56  0.56 -0.56  0.33 -0.56  0.56 -0.56]
  [ 0.11 -0.56  0.56 -0.78  0.56 -0.56  0.11]
  [-0.11  0.33 -0.78  1.   -0.78  0.33 -0.11]
  [ 0.11 -0.56  0.56 -0.78  0.56 -0.56  0.11]
  [-0.56  0.56 -0.56  0.33 -0.56  0.56 -0.56]
  [ 0.33 -0.56  0.11 -0.11  0.11 -0.56  0.33]]

 [[ 0.33 -0.11  0.56  0.33  0.11 -0.11  0.78]
  [-0.11  0.11 -0.11  0.33 -0.11  1.   -0.11]
  [ 0.56 -0.11  0.11 -0.33  1.   -0.11  0.11]
  [ 0.33  0.33 -0.33  0.56 -0.33  0.33  0.33]
  [ 0.11 -0.11  1.   -0.33  0.11 -0.11  0.56]
  [-0.11  1.   -0.11  0.33 -0.11  0.11 -0.11]
  [ 0.78 -0.11  0.11  0.33  0.56 -0.11  0.33]]]
pooling:
 [[1.   0.33 0.56 0.33]
 [0.33 1.   0.33 0.56]
 [0.56 0.33 1.   0.11]
 [0.33 0.56 0.11 0.78]]
pooling:
 [[0.56 0.33 0.56 0.33]
 [0.33 1.   0.56 0.11]
 [0.56 0.56 0.56 0.11]
 [0.33 0.11 0.11 0.33]]
pooling:
 [[0.33 0.56 1.   0.78]
 [0.56 0.56 1.   0.33]
 [1.   1.   0.11 0.56]
 [0.78 0.33 0.56 0.33]]
relu map :
 [[0.78 0.   0.11 0.33 0.56 0.   0.33]
 [0.   1.   0.   0.33 0.   0.11 0.  ]
 [0.11 0.   1.   0.   0.11 0.   0.56]
 [0.33 0.33 0.   0.56 0.   0.33 0.33]
 [0.56 0.   0.11 0.   1.   0.   0.11]
 [0.   0.11 0.   0.33 0.   1.   0.  ]
 [0.33 0.   0.56 0.33 0.11 0.   0.78]]
relu map :
 [[0.33 0.   0.11 0.   0.11 0.   0.33]
 [0.   0.56 0.   0.33 0.   0.56 0.  ]
 [0.11 0.   0.56 0.   0.56 0.   0.11]
 [0.   0.33 0.   1.   0.   0.33 0.  ]
 [0.11 0.   0.56 0.   0.56 0.   0.11]
 [0.   0.56 0.   0.33 0.   0.56 0.  ]
 [0.33 0.   0.11 0.   0.11 0.   0.33]]
relu map :
 [[0.33 0.   0.56 0.33 0.11 0.   0.78]
 [0.   0.11 0.   0.33 0.   1.   0.  ]
 [0.56 0.   0.11 0.   1.   0.   0.11]
 [0.33 0.33 0.   0.56 0.   0.33 0.33]
 [0.11 0.   1.   0.   0.11 0.   0.56]
 [0.   1.   0.   0.33 0.   0.11 0.  ]
 [0.78 0.   0.11 0.33 0.56 0.   0.33]]

2.Pytorch版本：调用函数实现 卷积-池化-激活

import numpy as np
import torch
import torch.nn as nn
 
x = torch.tensor([[[[-1, -1, -1, -1, -1, -1, -1, -1, -1],
                    [-1, 1, -1, -1, -1, -1, -1, 1, -1],
                    [-1, -1, 1, -1, -1, -1, 1, -1, -1],
                    [-1, -1, -1, 1, -1, 1, -1, -1, -1],
                    [-1, -1, -1, -1, 1, -1, -1, -1, -1],
                    [-1, -1, -1, 1, -1, 1, -1, -1, -1],
                    [-1, -1, 1, -1, -1, -1, 1, -1, -1],
                    [-1, 1, -1, -1, -1, -1, -1, 1, -1],
                    [-1, -1, -1, -1, -1, -1, -1, -1, -1]]]], dtype=torch.float)
print(x.shape)
print(x)
 
print("--------------- 卷积  ---------------")
conv1 = nn.Conv2d(1, 1, (3, 3), 1)  # in_channel , out_channel , kennel_size , stride
conv1.weight.data = torch.Tensor([[[[1, -1, -1],
                                    [-1, 1, -1],
                                    [-1, -1, 1]]
                                   ]])
conv2 = nn.Conv2d(1, 1, (3, 3), 1)  # in_channel , out_channel , kennel_size , stride
conv2.weight.data = torch.Tensor([[[[1, -1, 1],
                                    [-1, 1, -1],
                                    [1, -1, 1]]
                                   ]])
conv3 = nn.Conv2d(1, 1, (3, 3), 1)  # in_channel , out_channel , kennel_size , stride
conv3.weight.data = torch.Tensor([[[[-1, -1, 1],
                                    [-1, 1, -1],
                                    [1, -1, -1]]
                                   ]])
 
feature_map1 = conv1(x)
feature_map2 = conv2(x)
feature_map3 = conv3(x)
 
print(feature_map1 / 9)
print(feature_map2 / 9)
print(feature_map3 / 9)
 
print("--------------- 池化  ---------------")
max_pool = nn.MaxPool2d(2, padding=0, stride=2)  # Pooling
zeroPad = nn.ZeroPad2d(padding=(0, 1, 0, 1))  # pad 0 , Left Right Up Down
 
feature_map_pad_0_1 = zeroPad(feature_map1)
feature_pool_1 = max_pool(feature_map_pad_0_1)
feature_map_pad_0_2 = zeroPad(feature_map2)
feature_pool_2 = max_pool(feature_map_pad_0_2)
feature_map_pad_0_3 = zeroPad(feature_map3)
feature_pool_3 = max_pool(feature_map_pad_0_3)
 
print(feature_pool_1.size())
print(feature_pool_1 / 9)
print(feature_pool_2 / 9)
print(feature_pool_3 / 9)
 
print("--------------- 激活  ---------------")
activation_function = nn.ReLU()
 
feature_relu1 = activation_function(feature_map1)
feature_relu2 = activation_function(feature_map2)
feature_relu3 = activation_function(feature_map3)
print(feature_relu1 / 9)
print(feature_relu2 / 9)
print(feature_relu3 / 9)

实验结果：

torch.Size([1, 1, 9, 9])
tensor([[[[-1., -1., -1., -1., -1., -1., -1., -1., -1.],
          [-1.,  1., -1., -1., -1., -1., -1.,  1., -1.],
          [-1., -1.,  1., -1., -1., -1.,  1., -1., -1.],
          [-1., -1., -1.,  1., -1.,  1., -1., -1., -1.],
          [-1., -1., -1., -1.,  1., -1., -1., -1., -1.],
          [-1., -1., -1.,  1., -1.,  1., -1., -1., -1.],
          [-1., -1.,  1., -1., -1., -1.,  1., -1., -1.],
          [-1.,  1., -1., -1., -1., -1., -1.,  1., -1.],
          [-1., -1., -1., -1., -1., -1., -1., -1., -1.]]]])
--------------- 卷积  ---------------
tensor([[[[ 0.7642, -0.1247,  0.0975,  0.3197,  0.5419, -0.1247,  0.3197],
          [-0.1247,  0.9864, -0.1247,  0.3197, -0.1247,  0.0975, -0.1247],
          [ 0.0975, -0.1247,  0.9864, -0.3470,  0.0975, -0.1247,  0.5419],
          [ 0.3197,  0.3197, -0.3470,  0.5419, -0.3470,  0.3197,  0.3197],
          [ 0.5419, -0.1247,  0.0975, -0.3470,  0.9864, -0.1247,  0.0975],
          [-0.1247,  0.0975, -0.1247,  0.3197, -0.1247,  0.9864, -0.1247],
          [ 0.3197, -0.1247,  0.5419,  0.3197,  0.0975, -0.1247,  0.7642]]]],
       grad_fn=<DivBackward0>)
tensor([[[[ 0.3529, -0.5360,  0.1306, -0.0916,  0.1306, -0.5360,  0.3529],
          [-0.5360,  0.5751, -0.5360,  0.3529, -0.5360,  0.5751, -0.5360],
          [ 0.1306, -0.5360,  0.5751, -0.7582,  0.5751, -0.5360,  0.1306],
          [-0.0916,  0.3529, -0.7582,  1.0195, -0.7582,  0.3529, -0.0916],
          [ 0.1306, -0.5360,  0.5751, -0.7582,  0.5751, -0.5360,  0.1306],
          [-0.5360,  0.5751, -0.5360,  0.3529, -0.5360,  0.5751, -0.5360],
          [ 0.3529, -0.5360,  0.1306, -0.0916,  0.1306, -0.5360,  0.3529]]]],
       grad_fn=<DivBackward0>)
tensor([[[[ 0.3684, -0.0760,  0.5907,  0.3684,  0.1462, -0.0760,  0.8129],
          [-0.0760,  0.1462, -0.0760,  0.3684, -0.0760,  1.0351, -0.0760],
          [ 0.5907, -0.0760,  0.1462, -0.2982,  1.0351, -0.0760,  0.1462],
          [ 0.3684,  0.3684, -0.2982,  0.5907, -0.2982,  0.3684,  0.3684],
          [ 0.1462, -0.0760,  1.0351, -0.2982,  0.1462, -0.0760,  0.5907],
          [-0.0760,  1.0351, -0.0760,  0.3684, -0.0760,  0.1462, -0.0760],
          [ 0.8129, -0.0760,  0.1462,  0.3684,  0.5907, -0.0760,  0.3684]]]],
       grad_fn=<DivBackward0>)
--------------- 池化  ---------------
torch.Size([1, 1, 4, 4])
tensor([[[[0.9864, 0.3197, 0.5419, 0.3197],
          [0.3197, 0.9864, 0.3197, 0.5419],
          [0.5419, 0.3197, 0.9864, 0.0975],
          [0.3197, 0.5419, 0.0975, 0.7642]]]], grad_fn=<DivBackward0>)
tensor([[[[0.5751, 0.3529, 0.5751, 0.3529],
          [0.3529, 1.0195, 0.5751, 0.1306],
          [0.5751, 0.5751, 0.5751, 0.1306],
          [0.3529, 0.1306, 0.1306, 0.3529]]]], grad_fn=<DivBackward0>)
tensor([[[[0.3684, 0.5907, 1.0351, 0.8129],
          [0.5907, 0.5907, 1.0351, 0.3684],
          [1.0351, 1.0351, 0.1462, 0.5907],
          [0.8129, 0.3684, 0.5907, 0.3684]]]], grad_fn=<DivBackward0>)
--------------- 激活  ---------------
tensor([[[[0.7642, 0.0000, 0.0975, 0.3197, 0.5419, 0.0000, 0.3197],
          [0.0000, 0.9864, 0.0000, 0.3197, 0.0000, 0.0975, 0.0000],
          [0.0975, 0.0000, 0.9864, 0.0000, 0.0975, 0.0000, 0.5419],
          [0.3197, 0.3197, 0.0000, 0.5419, 0.0000, 0.3197, 0.3197],
          [0.5419, 0.0000, 0.0975, 0.0000, 0.9864, 0.0000, 0.0975],
          [0.0000, 0.0975, 0.0000, 0.3197, 0.0000, 0.9864, 0.0000],
          [0.3197, 0.0000, 0.5419, 0.3197, 0.0975, 0.0000, 0.7642]]]],
       grad_fn=<DivBackward0>)
tensor([[[[0.3529, 0.0000, 0.1306, 0.0000, 0.1306, 0.0000, 0.3529],
          [0.0000, 0.5751, 0.0000, 0.3529, 0.0000, 0.5751, 0.0000],
          [0.1306, 0.0000, 0.5751, 0.0000, 0.5751, 0.0000, 0.1306],
          [0.0000, 0.3529, 0.0000, 1.0195, 0.0000, 0.3529, 0.0000],
          [0.1306, 0.0000, 0.5751, 0.0000, 0.5751, 0.0000, 0.1306],
          [0.0000, 0.5751, 0.0000, 0.3529, 0.0000, 0.5751, 0.0000],
          [0.3529, 0.0000, 0.1306, 0.0000, 0.1306, 0.0000, 0.3529]]]],
       grad_fn=<DivBackward0>)
tensor([[[[0.3684, 0.0000, 0.5907, 0.3684, 0.1462, 0.0000, 0.8129],
          [0.0000, 0.1462, 0.0000, 0.3684, 0.0000, 1.0351, 0.0000],
          [0.5907, 0.0000, 0.1462, 0.0000, 1.0351, 0.0000, 0.1462],
          [0.3684, 0.3684, 0.0000, 0.5907, 0.0000, 0.3684, 0.3684],
          [0.1462, 0.0000, 1.0351, 0.0000, 0.1462, 0.0000, 0.5907],
          [0.0000, 1.0351, 0.0000, 0.3684, 0.0000, 0.1462, 0.0000],
          [0.8129, 0.0000, 0.1462, 0.3684, 0.5907, 0.0000, 0.3684]]]],
       grad_fn=<DivBackward0>)

3. 可视化：了解数字与图像之间的关系

import torch
import torch.nn as nn
import matplotlib.pyplot as plt

plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号 #有中文出现的情况，需要u'内容
x = torch.tensor([[[[-1, -1, -1, -1, -1, -1, -1, -1, -1],
                    [-1, 1, -1, -1, -1, -1, -1, 1, -1],
                    [-1, -1, 1, -1, -1, -1, 1, -1, -1],
                    [-1, -1, -1, 1, -1, 1, -1, -1, -1],
                    [-1, -1, -1, -1, 1, -1, -1, -1, -1],
                    [-1, -1, -1, 1, -1, 1, -1, -1, -1],
                    [-1, -1, 1, -1, -1, -1, 1, -1, -1],
                    [-1, 1, -1, -1, -1, -1, -1, 1, -1],
                    [-1, -1, -1, -1, -1, -1, -1, -1, -1]]]], dtype=torch.float)
print(x.shape)
print(x)
img = x.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('原图')
plt.show()

print("--------------- 卷积  ---------------")
conv1 = nn.Conv2d(1, 1, (3, 3), 1)  # in_channel , out_channel , kennel_size , stride
conv1.weight.data = torch.Tensor([[[[1, -1, -1],
                                    [-1, 1, -1],
                                    [-1, -1, 1]]
                                   ]])
img = conv1.weight.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('Kernel 1')
plt.show()
conv2 = nn.Conv2d(1, 1, (3, 3), 1)  # in_channel , out_channel , kennel_size , stride
conv2.weight.data = torch.Tensor([[[[1, -1, 1],
                                    [-1, 1, -1],
                                    [1, -1, 1]]
                                   ]])
img = conv2.weight.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('Kernel 2')
plt.show()
conv3 = nn.Conv2d(1, 1, (3, 3), 1)  # in_channel , out_channel , kennel_size , stride
conv3.weight.data = torch.Tensor([[[[-1, -1, 1],
                                    [-1, 1, -1],
                                    [1, -1, -1]]
                                   ]])
img = conv3.weight.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('Kernel 3')
plt.show()

feature_map1 = conv1(x)
feature_map2 = conv2(x)
feature_map3 = conv3(x)

print(feature_map1 / 9)
print(feature_map2 / 9)
print(feature_map3 / 9)

img = feature_map1.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('卷积后的特征图1')
plt.show()

print("--------------- 池化  ---------------")
max_pool = nn.MaxPool2d(2, padding=0, stride=2)  # Pooling
zeroPad = nn.ZeroPad2d(padding=(0, 1, 0, 1))  # pad 0 , Left Right Up Down

feature_map_pad_0_1 = zeroPad(feature_map1)
feature_pool_1 = max_pool(feature_map_pad_0_1)
feature_map_pad_0_2 = zeroPad(feature_map2)
feature_pool_2 = max_pool(feature_map_pad_0_2)
feature_map_pad_0_3 = zeroPad(feature_map3)
feature_pool_3 = max_pool(feature_map_pad_0_3)

print(feature_pool_1.size())
print(feature_pool_1 / 9)
print(feature_pool_2 / 9)
print(feature_pool_3 / 9)
img = feature_pool_1.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('卷积池化后的特征图1')
plt.show()

print("--------------- 激活  ---------------")
activation_function = nn.ReLU()

feature_relu1 = activation_function(feature_map1)
feature_relu2 = activation_function(feature_map2)
feature_relu3 = activation_function(feature_map3)
print(feature_relu1 / 9)
print(feature_relu2 / 9)
print(feature_relu3 / 9)
img = feature_relu1.data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.title('卷积 + relu 后的特征图1')
plt.show()

实验结果：

torch.Size([1, 1, 9, 9])
tensor([[[[-1., -1., -1., -1., -1., -1., -1., -1., -1.],
          [-1.,  1., -1., -1., -1., -1., -1.,  1., -1.],
          [-1., -1.,  1., -1., -1., -1.,  1., -1., -1.],
          [-1., -1., -1.,  1., -1.,  1., -1., -1., -1.],
          [-1., -1., -1., -1.,  1., -1., -1., -1., -1.],
          [-1., -1., -1.,  1., -1.,  1., -1., -1., -1.],
          [-1., -1.,  1., -1., -1., -1.,  1., -1., -1.],
          [-1.,  1., -1., -1., -1., -1., -1.,  1., -1.],
          [-1., -1., -1., -1., -1., -1., -1., -1., -1.]]]])
--------------- 卷积  ---------------
tensor([[[[ 0.7790, -0.1099,  0.1123,  0.3345,  0.5568, -0.1099,  0.3345],
          [-0.1099,  1.0012, -0.1099,  0.3345, -0.1099,  0.1123, -0.1099],
          [ 0.1123, -0.1099,  1.0012, -0.3321,  0.1123, -0.1099,  0.5568],
          [ 0.3345,  0.3345, -0.3321,  0.5568, -0.3321,  0.3345,  0.3345],
          [ 0.5568, -0.1099,  0.1123, -0.3321,  1.0012, -0.1099,  0.1123],
          [-0.1099,  0.1123, -0.1099,  0.3345, -0.1099,  1.0012, -0.1099],
          [ 0.3345, -0.1099,  0.5568,  0.3345,  0.1123, -0.1099,  0.7790]]]],
       grad_fn=<DivBackward0>)
tensor([[[[ 0.3603, -0.5286,  0.1381, -0.0841,  0.1381, -0.5286,  0.3603],
          [-0.5286,  0.5825, -0.5286,  0.3603, -0.5286,  0.5825, -0.5286],
          [ 0.1381, -0.5286,  0.5825, -0.7508,  0.5825, -0.5286,  0.1381],
          [-0.0841,  0.3603, -0.7508,  1.0270, -0.7508,  0.3603, -0.0841],
          [ 0.1381, -0.5286,  0.5825, -0.7508,  0.5825, -0.5286,  0.1381],
          [-0.5286,  0.5825, -0.5286,  0.3603, -0.5286,  0.5825, -0.5286],
          [ 0.3603, -0.5286,  0.1381, -0.0841,  0.1381, -0.5286,  0.3603]]]],
       grad_fn=<DivBackward0>)
tensor([[[[ 0.3475, -0.0969,  0.5698,  0.3475,  0.1253, -0.0969,  0.7920],
          [-0.0969,  0.1253, -0.0969,  0.3475, -0.0969,  1.0142, -0.0969],
          [ 0.5698, -0.0969,  0.1253, -0.3191,  1.0142, -0.0969,  0.1253],
          [ 0.3475,  0.3475, -0.3191,  0.5698, -0.3191,  0.3475,  0.3475],
          [ 0.1253, -0.0969,  1.0142, -0.3191,  0.1253, -0.0969,  0.5698],
          [-0.0969,  1.0142, -0.0969,  0.3475, -0.0969,  0.1253, -0.0969],
          [ 0.7920, -0.0969,  0.1253,  0.3475,  0.5698, -0.0969,  0.3475]]]],
       grad_fn=<DivBackward0>)
--------------- 池化  ---------------
torch.Size([1, 1, 4, 4])
tensor([[[[1.0012, 0.3345, 0.5568, 0.3345],
          [0.3345, 1.0012, 0.3345, 0.5568],
          [0.5568, 0.3345, 1.0012, 0.1123],
          [0.3345, 0.5568, 0.1123, 0.7790]]]], grad_fn=<DivBackward0>)
tensor([[[[0.5825, 0.3603, 0.5825, 0.3603],
          [0.3603, 1.0270, 0.5825, 0.1381],
          [0.5825, 0.5825, 0.5825, 0.1381],
          [0.3603, 0.1381, 0.1381, 0.3603]]]], grad_fn=<DivBackward0>)
tensor([[[[0.3475, 0.5698, 1.0142, 0.7920],
          [0.5698, 0.5698, 1.0142, 0.3475],
          [1.0142, 1.0142, 0.1253, 0.5698],
          [0.7920, 0.3475, 0.5698, 0.3475]]]], grad_fn=<DivBackward0>)
--------------- 激活  ---------------
tensor([[[[0.7790, 0.0000, 0.1123, 0.3345, 0.5568, 0.0000, 0.3345],
          [0.0000, 1.0012, 0.0000, 0.3345, 0.0000, 0.1123, 0.0000],
          [0.1123, 0.0000, 1.0012, 0.0000, 0.1123, 0.0000, 0.5568],
          [0.3345, 0.3345, 0.0000, 0.5568, 0.0000, 0.3345, 0.3345],
          [0.5568, 0.0000, 0.1123, 0.0000, 1.0012, 0.0000, 0.1123],
          [0.0000, 0.1123, 0.0000, 0.3345, 0.0000, 1.0012, 0.0000],
          [0.3345, 0.0000, 0.5568, 0.3345, 0.1123, 0.0000, 0.7790]]]],
       grad_fn=<DivBackward0>)
tensor([[[[0.3603, 0.0000, 0.1381, 0.0000, 0.1381, 0.0000, 0.3603],
          [0.0000, 0.5825, 0.0000, 0.3603, 0.0000, 0.5825, 0.0000],
          [0.1381, 0.0000, 0.5825, 0.0000, 0.5825, 0.0000, 0.1381],
          [0.0000, 0.3603, 0.0000, 1.0270, 0.0000, 0.3603, 0.0000],
          [0.1381, 0.0000, 0.5825, 0.0000, 0.5825, 0.0000, 0.1381],
          [0.0000, 0.5825, 0.0000, 0.3603, 0.0000, 0.5825, 0.0000],
          [0.3603, 0.0000, 0.1381, 0.0000, 0.1381, 0.0000, 0.3603]]]],
       grad_fn=<DivBackward0>)
tensor([[[[0.3475, 0.0000, 0.5698, 0.3475, 0.1253, 0.0000, 0.7920],
          [0.0000, 0.1253, 0.0000, 0.3475, 0.0000, 1.0142, 0.0000],
          [0.5698, 0.0000, 0.1253, 0.0000, 1.0142, 0.0000, 0.1253],
          [0.3475, 0.3475, 0.0000, 0.5698, 0.0000, 0.3475, 0.3475],
          [0.1253, 0.0000, 1.0142, 0.0000, 0.1253, 0.0000, 0.5698],
          [0.0000, 1.0142, 0.0000, 0.3475, 0.0000, 0.1253, 0.0000],
          [0.7920, 0.0000, 0.1253, 0.3475, 0.5698, 0.0000, 0.3475]]]],
       grad_fn=<DivBackward0>)

二、 基于CNN的XO识别

1. 数据集
共2000张图片，X、O各1000张。
从X、O文件夹，分别取出150张作为测试集。
文件夹train_data：放置训练集 1700张图片，文件夹test_data： 放置测试集 300张图片

2. 构建模型

import torch.nn as nn
import torch
class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 9, 3)
        self.maxpool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(9, 5, 3)
 
        self.relu = nn.ReLU()
        self.fc1 = nn.Linear(27 * 27 * 5, 1200)
        self.fc2 = nn.Linear(1200, 64)
        self.fc3 = nn.Linear(64, 2)
 
    def forward(self, x):
        x = self.maxpool(self.relu(self.conv1(x)))
        x = self.maxpool(self.relu(self.conv2(x)))
        x = x.view(-1, 27 * 27 * 5)
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)
        return x

3. 训练模型

import torch.nn as nn
import torch


class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 9, 3)
        self.maxpool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(9, 5, 3)

        self.relu = nn.ReLU()
        self.fc1 = nn.Linear(27 * 27 * 5, 1200)
        self.fc2 = nn.Linear(1200, 64)
        self.fc3 = nn.Linear(64, 2)

    def forward(self, x):
        x = self.maxpool(self.relu(self.conv1(x)))
        x = self.maxpool(self.relu(self.conv2(x)))
        x = x.view(-1, 27 * 27 * 5)
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)
        return x


import torch
from torchvision import transforms, datasets
import torch.nn as nn
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import torch.optim as optim

transforms = transforms.Compose([
    transforms.ToTensor(),  # 把图片进行归一化，并把数据转换成Tensor类型
    transforms.Grayscale(1)  # 把图片 转为灰度图
])

path = r'C:\Users\86181\Desktop\1\train_data'
path_test = r'C:\Users\86181\Desktop\1\test_data'

data_train = datasets.ImageFolder(path, transform=transforms)
data_test = datasets.ImageFolder(path_test, transform=transforms)

print("size of train_data:", len(data_train))
print("size of test_data:", len(data_test))

data_loader = DataLoader(data_train, batch_size=64, shuffle=True)
data_loader_test = DataLoader(data_test, batch_size=64, shuffle=True)
model = Net()

criterion = torch.nn.CrossEntropyLoss()  # 损失函数 交叉熵损失函数
optimizer = torch.optim.SGD(model.parameters(), lr=0.1)  # 优化函数：随机梯度下降

epochs = 10
for epoch in range(epochs):
    running_loss = 0.0
    for i, data in enumerate(data_loader):
        images, label = data
        out = model(images)
        loss = criterion(out, label)

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        running_loss += loss.item()
        if (i + 1) % 10 == 0:
            print('[%d  %5d]   loss: %.6f' % (epoch + 1, i + 1, running_loss / 100))
            running_loss = 0.0

print('finished train')

# 保存模型
torch.save(model, 'model_name.pth')  # 保存的是模型， 不止是w和b权重值

实验结果：

size of train_data: 1700
size of test_data: 300
[1     10]   loss: 0.069
[1     20]   loss: 0.069
[2     10]   loss: 0.068
[2     20]   loss: 0.066
[3     10]   loss: 0.059
[3     20]   loss: 0.036
[4     10]   loss: 0.050
[4     20]   loss: 0.018
[5     10]   loss: 0.012
[5     20]   loss: 0.005
[6     10]   loss: 0.004
[6     20]   loss: 0.003
[7     10]   loss: 0.001
[7     20]   loss: 0.003
[8     10]   loss: 0.001
[8     20]   loss: 0.001
[9     10]   loss: 0.001
[9     20]   loss: 0.000
[10     10]   loss: 0.000
[10     20]   loss: 0.001
finished train

4. 测试训练好的模型

# 读取模型
model_load = torch.load('model_name.pth')
# 读取一张图片 images[0]，测试
print("labels[0] truth:\t", labels[0])
x = images[0]
predicted = torch.max(model_load(x), 1)
print("labels[0] predict:\t", predicted.indices)
 
img = images[0].data.squeeze().numpy()  # 将输出转换为图片的格式
plt.imshow(img, cmap='gray')
plt.show()

实验结果：

labels[0] truth:	 tensor(0)
labels[0] predict:	 tensor([0])

5. 计算模型的准确率

model = Net()
model.load_state_dict(torch.load('model_name1.pth', map_location='cpu'))  # 导入网络的参数
 
correct = 0
total = 0
with torch.no_grad():  # 进行评测的时候网络不更新梯度
    for data in data_loader_test:  # 读取测试集
        images, labels = data
        outputs = model(images)
        _, predicted = torch.max(outputs.data, 1)  # 取出 最大值的索引 作为 分类结果
        total += labels.size(0)  # labels 的长度
        correct += (predicted == labels).sum().item()  # 预测正确的数目
print('Accuracy of the network on the  test images: %f %%' % (100. * correct / total))

实验结果：

Accuracy of the network on the  test images: 99.333333 %

6. 查看训练好的模型的特征图

import torch
import matplotlib.pyplot as plt
import numpy as np
from PIL import Image
from torchvision import transforms, datasets
import torch.nn as nn
from torch.utils.data import DataLoader
 
#  定义图像预处理过程(要与网络模型训练过程中的预处理过程一致)
 
transforms = transforms.Compose([
    transforms.ToTensor(),  # 把图片进行归一化，并把数据转换成Tensor类型
    transforms.Grayscale(1)  # 把图片 转为灰度图
])
path = r'C:\Users\86181\Desktop\1\train_data'
data_train = datasets.ImageFolder(path, transform=transforms)
data_loader = DataLoader(data_train, batch_size=64, shuffle=True)
for i, data in enumerate(data_loader):
    images, labels = data
    print(images.shape)
    print(labels.shape)
    break
 
 
class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 9, 3)  # in_channel , out_channel , kennel_size , stride
        self.maxpool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(9, 5, 3)  # in_channel , out_channel , kennel_size , stride
 
        self.relu = nn.ReLU()
        self.fc1 = nn.Linear(27 * 27 * 5, 1200)  # full connect 1
        self.fc2 = nn.Linear(1200, 64)  # full connect 2
        self.fc3 = nn.Linear(64, 2)  # full connect 3
 
    def forward(self, x):
        outputs = []
        x = self.conv1(x)
        outputs.append(x)
        x = self.relu(x)
        outputs.append(x)
        x = self.maxpool(x)
        outputs.append(x)
        x = self.conv2(x)
 
        x = self.relu(x)
 
        x = self.maxpool(x)
 
        x = x.view(-1, 27 * 27 * 5)
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)
        return outputs
 
 
# create model
model1 = Net()
 
# load model weights加载预训练权重
# model_weight_path ="./AlexNet.pth"
model_weight_path = "model_name1.pth"
model1.load_state_dict(torch.load(model_weight_path))
 
# 打印出模型的结构
print(model1)
 
x = images[0]
x = x.reshape([1, 1, 116, 116])
 
# forward正向传播过程
out_put = model1(x)
 
# forward正向传播过程
for feature_map in out_put:
    # [N, C, H, W] -> [C, H, W]    维度变换
    im = np.squeeze(feature_map.detach().numpy())
    # [C, H, W] -> [H, W, C]
    im = np.transpose(im, [1, 2, 0])
    print(im.shape)
 
    # show 9 feature maps
    plt.figure()
    for i in range(9):
        ax = plt.subplot(3, 3, i + 1)  # 参数意义：3：图片绘制行数，5：绘制图片列数，i+1：图的索引
        # [H, W, C]
        # 特征矩阵每一个channel对应的是一个二维的特征矩阵，就像灰度图像一样，channel=1
        # plt.imshow(im[:, :, i])
        plt.imshow(im[:, :, i], cmap='gray')
    plt.show()

实验结果：

Net(
  (conv1): Conv2d(1, 9, kernel_size=(3, 3), stride=(1, 1))
  (maxpool): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)
  (conv2): Conv2d(9, 5, kernel_size=(3, 3), stride=(1, 1))
  (relu): ReLU()
  (fc1): Linear(in_features=3645, out_features=1200, bias=True)
  (fc2): Linear(in_features=1200, out_features=64, bias=True)
  (fc3): Linear(in_features=64, out_features=2, bias=True)
)
(114, 114, 9)
(114, 114, 9)
(57, 57, 9)

7. 查看训练好的模型的卷积核

# 看看每层的 卷积核 长相，特征图 长相
# 获取网络结构的特征矩阵并可视化
import torch
import matplotlib.pyplot as plt
import numpy as np
from PIL import Image
from torchvision import transforms, datasets
import torch.nn as nn
from torch.utils.data import DataLoader
 
plt.rcParams['font.sans-serif'] = ['SimHei']  # 用来正常显示中文标签
plt.rcParams['axes.unicode_minus'] = False  # 用来正常显示负号 #有中文出现的情况，需要u'内容
#  定义图像预处理过程(要与网络模型训练过程中的预处理过程一致)
transforms = transforms.Compose([
    transforms.ToTensor(),  # 把图片进行归一化，并把数据转换成Tensor类型
    transforms.Grayscale(1)  # 把图片 转为灰度图
])
path = r'C:\Users\86181\Desktop\1\train_data'
data_train = datasets.ImageFolder(path, transform=transforms)
data_loader = DataLoader(data_train, batch_size=64, shuffle=True)
for i, data in enumerate(data_loader):
    images, labels = data
    # print(images.shape)
    # print(labels.shape)
    break
 
 
class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 9, 3)  # in_channel , out_channel , kennel_size , stride
        self.maxpool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(9, 5, 3)  # in_channel , out_channel , kennel_size , stride
 
        self.relu = nn.ReLU()
        self.fc1 = nn.Linear(27 * 27 * 5, 1200)  # full connect 1
        self.fc2 = nn.Linear(1200, 64)  # full connect 2
        self.fc3 = nn.Linear(64, 2)  # full connect 3
 
    def forward(self, x):
        outputs = []
        x = self.maxpool(self.relu(self.conv1(x)))
        # outputs.append(x)
        x = self.maxpool(self.relu(self.conv2(x)))
        outputs.append(x)
        x = x.view(-1, 27 * 27 * 5)
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)
        return outputs
 
 
# create model
model1 = Net()
 
# load model weights加载预训练权重
model_weight_path = "model_name1.pth"
model1.load_state_dict(torch.load(model_weight_path))
 
x = images[0]
x = x.reshape([1, 1, 116, 116])
# forward正向传播过程
out_put = model1(x)
 
weights_keys = model1.state_dict().keys()
for key in weights_keys:
    print("key :", key)
    # 卷积核通道排列顺序 [kernel_number, kernel_channel, kernel_height, kernel_width]
    if key == "conv1.weight":
        weight_t = model1.state_dict()[key].numpy()
        print("weight_t.shape", weight_t.shape)
        k = weight_t[:, 0, :, :]  # 获取第一个卷积核的信息参数
        # show 9 kernel ,1 channel
        plt.figure()
 
        for i in range(9):
            ax = plt.subplot(3, 3, i + 1)  # 参数意义：3：图片绘制行数，5：绘制图片列数，i+1：图的索引
            plt.imshow(k[i, :, :], cmap='gray')
            title_name = 'kernel' + str(i) + ',channel1'
            plt.title(title_name)
        plt.show()
 
    if key == "conv2.weight":
        weight_t = model1.state_dict()[key].numpy()
        print("weight_t.shape", weight_t.shape)
        k = weight_t[:, :, :, :]  # 获取第一个卷积核的信息参数
        print(k.shape)
        print(k)
 
        plt.figure()
        for c in range(9):
            channel = k[:, c, :, :]
            for i in range(5):
                ax = plt.subplot(2, 3, i + 1)  # 参数意义：3：图片绘制行数，5：绘制图片列数，i+1：图的索引
                plt.imshow(channel[i, :, :], cmap='gray')
                title_name = 'kernel' + str(i) + ',channel' + str(c)
                plt.title(title_name)
            plt.show()

8. 训练模型源代码

import torch.nn as nn
import torch


class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 9, 3)
        self.maxpool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(9, 5, 3)

        self.relu = nn.ReLU()
        self.fc1 = nn.Linear(27 * 27 * 5, 1200)
        self.fc2 = nn.Linear(1200, 64)
        self.fc3 = nn.Linear(64, 2)

    def forward(self, x):
        x = self.maxpool(self.relu(self.conv1(x)))
        x = self.maxpool(self.relu(self.conv2(x)))
        x = x.view(-1, 27 * 27 * 5)
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)
        return x


import torch
from torchvision import transforms, datasets
import torch.nn as nn
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import torch.optim as optim

transforms = transforms.Compose([
    transforms.ToTensor(),  # 把图片进行归一化，并把数据转换成Tensor类型
    transforms.Grayscale(1)  # 把图片 转为灰度图
])

path = r'C:\Users\86181\Desktop\1\train_data'
path_test = r'C:\Users\86181\Desktop\1\test_data'

data_train = datasets.ImageFolder(path, transform=transforms)
data_test = datasets.ImageFolder(path_test, transform=transforms)

print("size of train_data:", len(data_train))
print("size of test_data:", len(data_test))

data_loader = DataLoader(data_train, batch_size=64, shuffle=True)
data_loader_test = DataLoader(data_test, batch_size=64, shuffle=True)
model = Net()

criterion = torch.nn.CrossEntropyLoss()  # 损失函数 交叉熵损失函数
optimizer = torch.optim.SGD(model.parameters(), lr=0.1)  # 优化函数：随机梯度下降

epochs = 10
for epoch in range(epochs):
    running_loss = 0.0
    for i, data in enumerate(data_loader):
        images, label = data
        out = model(images)
        loss = criterion(out, label)
        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

        running_loss += loss.item()
        if (i + 1) % 10 == 0:
            print('[%d  %5d]   loss: %.6f' % (epoch + 1, i + 1, running_loss / 100))
            running_loss = 0.0

print('finished train')

# 保存模型
torch.save(model, 'model_name.pth')  # 保存的是模型， 不止是w和b权重值

9. 测试模型源代码

# https://blog.csdn.net/qq_53345829/article/details/124308515
import torch
from torchvision import transforms, datasets
import torch.nn as nn
from torch.utils.data import DataLoader
import matplotlib.pyplot as plt
import torch.optim as optim
 
transforms = transforms.Compose([
    transforms.ToTensor(),  # 把图片进行归一化，并把数据转换成Tensor类型
    transforms.Grayscale(1)  # 把图片 转为灰度图
])
 
path = r'C:\Users\86181\Desktop\1\train_data'
path_test = r'C:\Users\86181\Desktop\1\test_data'
 
data_train = datasets.ImageFolder(path, transform=transforms)
data_test = datasets.ImageFolder(path_test, transform=transforms)
 
print("size of train_data:", len(data_train))
print("size of test_data:", len(data_test))
 
data_loader = DataLoader(data_train, batch_size=64, shuffle=True)
data_loader_test = DataLoader(data_test, batch_size=64, shuffle=True)
print(len(data_loader))
print(len(data_loader_test))
 
 
class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 9, 3)  # in_channel , out_channel , kennel_size , stride
        self.maxpool = nn.MaxPool2d(2, 2)
        self.conv2 = nn.Conv2d(9, 5, 3)  # in_channel , out_channel , kennel_size , stride
 
        self.relu = nn.ReLU()
        self.fc1 = nn.Linear(27 * 27 * 5, 1200)  # full connect 1
        self.fc2 = nn.Linear(1200, 64)  # full connect 2
        self.fc3 = nn.Linear(64, 2)  # full connect 3
 
    def forward(self, x):
        x = self.maxpool(self.relu(self.conv1(x)))
        x = self.maxpool(self.relu(self.conv2(x)))
        x = x.view(-1, 27 * 27 * 5)
        x = self.relu(self.fc1(x))
        x = self.relu(self.fc2(x))
        x = self.fc3(x)
        return x
 
# 读取模型
model = Net()
model.load_state_dict(torch.load('model_name1.pth', map_location='cpu')) # 导入网络的参数
 
# model_load = torch.load('model_name1.pth')
# https://blog.csdn.net/qq_41360787/article/details/104332706
 
correct = 0
total = 0
with torch.no_grad():  # 进行评测的时候网络不更新梯度
    for data in data_loader_test:  # 读取测试集
        images, labels = data
        outputs = model(images)
        _, predicted = torch.max(outputs.data, 1)  # 取出 最大值的索引 作为 分类结果
        total += labels.size(0)  # labels 的长度
        correct += (predicted == labels).sum().item()  # 预测正确的数目
print('Accuracy of the network on the  test images: %f %%' % (100. * correct / total))

实验心得：这次试验，使用Numpy手写底层代码和使用框架进行卷积池化操作，体会到了卷积与池化的主要用处，理解卷积过程中的卷积核由来、特征图由来以及卷积核参数的设置，收获许多。