这个代码实现了预测和可视化
1 import os 2 3 # third-party library 4 import torch 5 import torch.nn as nn 6 import torch.utils.data as Data 7 import torchvision 8 import matplotlib.pyplot as plt 9 10 # torch.manual_seed(1) # reproducible 11 12 # Hyper Parameters 13 EPOCH = 1 # train the training data n times, to save time, we just train 1 epoch 14 BATCH_SIZE = 50 15 LR = 0.001 # learning rate 16 DOWNLOAD_MNIST = False 17 18 19 # Mnist digits dataset 20 if not(os.path.exists('./mnist/')) or not os.listdir('./mnist/'): 21 # not mnist dir or mnist is empyt dir 22 DOWNLOAD_MNIST = True 23 24 train_data = torchvision.datasets.MNIST( 25 root='./mnist/', 26 train=True, # this is training data 27 transform=torchvision.transforms.ToTensor(), # 把数据压缩到0到1之间的numpy数据 28 # torch.FloatTensor of shape (C x H x W) and normalize in the range [0.0, 1.0] 29 download=DOWNLOAD_MNIST, 30 ) 31 32 # plot one example 33 print(train_data.train_data.size()) # (60000, 28, 28) 34 print(train_data.train_labels.size()) # (60000) 35 plt.imshow(train_data.train_data[0].numpy(), cmap='gray') 36 plt.title('%i' % train_data.train_labels[0]) 37 plt.show() 38 39 # Data Loader for easy mini-batch return in training, the image batch shape will be (50, 1, 28, 28) 40 train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True) 41 42 # pick 2000 samples to speed up testing 43 test_data = torchvision.datasets.MNIST(root='./mnist/', train=False) 44 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) 45 test_y = test_data.test_labels[:2000] 46 47 48 class CNN(nn.Module): 49 def __init__(self): 50 super(CNN, self).__init__() 51 self.conv1 = nn.Sequential( # input shape (1, 28, 28) 52 nn.Conv2d( 53 in_channels=1, # input height 54 out_channels=16, # n_filters 55 kernel_size=5, # filter size 56 stride=1, # filter movement/step 57 padding=2, # if want same width and length of this image after Conv2d, padding=(kernel_size-1)/2 if stride=1 58 ), # output shape (16, 28, 28) 59 nn.ReLU(), # activation 60 nn.MaxPool2d(kernel_size=2), # choose max value in 2x2 area, output shape (16, 14, 14) 61 ) 62 self.conv2 = nn.Sequential( # input shape (16, 14, 14) 63 nn.Conv2d(16, 32, 5, 1, 2), # output shape (32, 14, 14) 64 nn.ReLU(), # activation 65 nn.MaxPool2d(2), # output shape (32, 7, 7) 66 ) 67 self.out = nn.Linear(32 * 7 * 7, 10) # fully connected layer, output 10 classes 68 69 def forward(self, x): 70 x = self.conv1(x) 71 x = self.conv2(x) 72 x = x.view(x.size(0), -1) # flatten the output of conv2 to (batch_size, 32 * 7 * 7) 73 output = self.out(x) 74 return output, x # return x for visualization 75 76 77 cnn = CNN() 78 print(cnn) # net architecture 79 80 optimizer = torch.optim.Adam(cnn.parameters(), lr=LR) # optimize all cnn parameters 81 loss_func = nn.CrossEntropyLoss() # the target label is not one-hotted 82 83 # following function (plot_with_labels) is for visualization, can be ignored if not interested 84 from matplotlib import cm 85 try: from sklearn.manifold import TSNE; HAS_SK = True 86 except: HAS_SK = False; print('Please install sklearn for layer visualization') 87 def plot_with_labels(lowDWeights, labels): 88 plt.cla() 89 X, Y = lowDWeights[:, 0], lowDWeights[:, 1] 90 for x, y, s in zip(X, Y, labels): 91 c = cm.rainbow(int(255 * s / 9)); plt.text(x, y, s, backgroundcolor=c, fontsize=9) 92 plt.xlim(X.min(), X.max()); plt.ylim(Y.min(), Y.max()); plt.title('Visualize last layer'); plt.show(); plt.pause(0.01) 93 94 plt.ion() 95 # training and testing 96 for epoch in range(EPOCH): 97 for step, (b_x, b_y) in enumerate(train_loader): # gives batch data, normalize x when iterate train_loader 98 99 output = cnn(b_x)[0] # cnn output 100 loss = loss_func(output, b_y) # cross entropy loss 101 optimizer.zero_grad() # clear gradients for this training step 102 loss.backward() # backpropagation, compute gradients 103 optimizer.step() # apply gradients 104 105 if step % 50 == 0: 106 test_output, last_layer = cnn(test_x) 107 pred_y = torch.max(test_output, 1)[1].data.numpy() 108 accuracy = float((pred_y == test_y.data.numpy()).astype(int).sum()) / float(test_y.size(0)) 109 print('Epoch: ', epoch, '| train loss: %.4f' % loss.data.numpy(), '| test accuracy: %.2f' % accuracy) 110 if HAS_SK: 111 # Visualization of trained flatten layer (T-SNE) 112 tsne = TSNE(perplexity=30, n_components=2, init='pca', n_iter=5000) 113 plot_only = 500 114 low_dim_embs = tsne.fit_transform(last_layer.data.numpy()[:plot_only, :]) 115 labels = test_y.numpy()[:plot_only] 116 plot_with_labels(low_dim_embs, labels) 117 plt.ioff() 118 119 # print 10 predictions from test data 120 test_output, _ = cnn(test_x[:10]) 121 pred_y = torch.max(test_output, 1)[1].data.numpy() 122 print(pred_y, 'prediction number') 123 print(test_y[:10].numpy(), 'real number')
去掉可视化进行代码简化
1 import os 2 # third-party library 3 import torch 4 import torch.nn as nn 5 import torch.utils.data as Data 6 import torchvision 7 import matplotlib.pyplot as plt 8 9 EPOCH = 1 # train the training data n times, to save time, we just train 1 epoch 10 BATCH_SIZE = 50 11 LR = 0.001 # learning rate 12 13 train_data = torchvision.datasets.MNIST( 14 root='./mnist/', #下载后的存放目录 15 train=True, # this is training data 16 transform=torchvision.transforms.ToTensor(), # 把数据压缩到0到1之间的numpy数据，如果原始数据是rgb数据（0-255）则变为黑白数据，并使numpy数据变为tensor数据 # torch.FloatTensor of shape (C x H x W) and normalize in the range [0.0, 1.0] 17 download=True#不存在该数据就设置为True进行下载，存在则改为False 18 ) 19 20 # plot one example 21 print(train_data.train_data.size()) # (60000, 28, 28)，六万图片 22 print(train_data.train_labels.size()) # (60000)，六万标签 23 plt.imshow(train_data.train_data[0].numpy(), cmap='gray')#展现第一个训练数据图片 24 plt.title('%i' % train_data.train_labels[0]) 25 plt.show() 26 27 # Data Loader for easy mini-batch return in training, the image batch shape will be (50, 1, 28, 28) 28 train_loader = Data.DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True) 29 30 # pick 2000 samples to speed up testing 31 test_data = torchvision.datasets.MNIST(root='./mnist/', train=False) 32 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) 33 test_y = test_data.test_labels[:2000] 34 35 class CNN(nn.Module): 36 def __init__(self): 37 super(CNN, self).__init__() 38 self.conv1 = nn.Sequential( # input shape (1, 28, 28)，考虑batch是（batch，1,28,28） 39 nn.Conv2d( 40 in_channels=1, # input height 41 out_channels=16, # n_filters 42 kernel_size=5, # filter size 43 stride=1, # filter movement/step 44 padding=2, # if want same width and length of this image after Conv2d, padding=(kernel_size-1)/2 if stride=1 45 ), # output shape (16, 28, 28) 46 nn.ReLU(), # activation 47 nn.MaxPool2d(kernel_size=2), # choose max value in 2x2 area, output shape (16, 14, 14) 48 ) 49 self.conv2 = nn.Sequential( # input shape (16, 14, 14) 50 nn.Conv2d(16, 32, 5, 1, 2), # output shape (32, 14, 14) 51 nn.ReLU(), # activation 52 nn.MaxPool2d(2), # output shape (32, 7, 7) 53 ) 54 self.out = nn.Linear(32 * 7 * 7, 10) # fully connected layer, output 10 classes 55 56 def forward(self, x): 57 x = self.conv1(x) 58 x = self.conv2(x) 59 x = x.view(x.size(0), -1) # flatten the output of conv2 to (batch_size, 32 * 7 * 7)，只有tensor对象才可以使用x.size(0) 60 output = self.out(x) 61 return output, x # return x for visualization 62 63 64 cnn = CNN() 65 #print(cnn) # net architecture 66 67 optimizer = torch.optim.Adam(cnn.parameters(), lr=LR) # optimize all cnn parameters 68 loss_func = nn.CrossEntropyLoss() # the target label is not one-hotted 69 70 # training and testing 71 for epoch in range(EPOCH): 72 for step, (b_x, b_y) in enumerate(train_loader): # gives batch data, normalize x when iterate train_loader 73 74 output = cnn(b_x)[0] # cnn output 75 loss = loss_func(output, b_y) # cross entropy loss 76 optimizer.zero_grad() # clear gradients for this training step 77 loss.backward() # backpropagation, compute gradients 78 optimizer.step() # apply gradients 79 80 if step % 50 == 0: 81 test_output = cnn(test_x)[0] 82 print("----------------") 83 #print(test_output.shape) #2000*10 84 #print(torch.max(test_output, 1)) #返回的每一行中最大值和其下标 85 pred_y = torch.max(test_output, 1)[1].data.numpy() #返回的是每个样本对应0-9数字可能性最大的概率对应的下标 86 accuracy = float((pred_y == test_y.data.numpy()).astype(int).sum()) / float(test_y.size(0)) 87 print('Epoch: ', epoch, '| train loss: %.4f' % loss.data.numpy(), '| test accuracy: %.2f' % accuracy) 88 89 # print 10 predictions from test data 90 test_output, _ = cnn(test_x[:10]) 91 pred_y = torch.max(test_output, 1)[1].data.numpy() 92 print(pred_y, 'prediction number') 93 print(test_y[:10].numpy(), 'real number')