本篇文章是记录使用简单神经网络进行对手写数字辨识
代码如下(示例):
可视化loss下降曲线
import torch
from matplotlib import pyplot as plt
#loss curve
def plot_curve(data):
fig = plt.figure()
plt.plot(range(len(data)), data, color = 'blue')
plt.legend(['value'], loc = 'upper right') #graphic symbol
plt.xlabel('step')
plt.ylabel('value')
plt.show()
可视化识别结果
# visualization of result
def plot_image(image, label, name):
fig = plt.figure()
for i in range(6):
plt.subplot(2, 3, i+1)
plt.tight_layout()
plt.imshow(img[i][0]*0.3081+0.1307, cmap='gray', interpolation='none')
plt.title("{0}: {1}".format(name, label[i].item()))
plt.xticks([])
plt.yticks([])
plt.show()
对标签进行one-hot编码
def one_hot(label, depth=10):
out = torch.zeros(label.size(0), depth)
idx = torch.LongTensor(label).view(-1,1)
out.scatter_(dim=1, index=idx, value=1)
return out
代码如下(示例):
import torch
from torch import nn
from torch.nn import functional as F
from torch import optim
import torchvision
from matplotlib import pyplot as plt
from utils import plot_image, plot_curve, one_hot
batch_size选择256,需要对训练数据进行随机打散,测试集不需要。
batch_size = 256
#load dataset
train_loader = torch.utils.data.DataLoader(
torchvision.datasets.MNIST('mnist_data', train=True, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize(
(0.1307,), (0.3081)) #mean=0.1307,sd=0.3081
])),
batch_size=batch_size, shuffle=True)
test_loader = torch.utils.data.DataLoader(
torchvision.datasets.MNIST('mnist_data/', train=False, download=True,
transform=torchvision.transforms.Compose([
torchvision.transforms.ToTensor(),
torchvision.transforms.Normalize(
(0.1307,), (0.3081)) #mean=0.1307,sd=0.3081
])),
batch_size=batch_size, shuffle=False)
可以看一下我们数据中的一个sample
x, y = next(iter(train_loader))
print(x.shape, y.shape, x.min(),x.max())
plot_image(x, y, 'sample')
我们写一个简单的神经网络,对一般的线性回归进行3层嵌套,也就是对y = w*x + b 进行3层嵌套,每一层再加一个ReLu激活函数,代码如下所示:
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
# xw+b
self.fc1 = nn.Linear(28*28, 256)
self.fc2 = nn.Linear(256, 64)
self.fc3 = nn.Linear(64, 10)
def forward(self, x):
# x:[b,1,28,28]
# h1 = relu(xw1+b)
x = F.relu(self.fc1(x))
# h2 = relu(h1w2 + b)
x = F.relu(self.fc2(x))
# h3 = h2w3 + b
x = self.fc3(x)
return x
接下来对神经网络进行训练(3次),并通过SGDM求解最优解,设置learning rate = 0.01,动量momentum=0.9,loss函数为MSE
net = Net()
optimizer = optim.SGD(net.parameters(), lr=0.01, momentum=0.9)
train_loss = []
for epoch in range(3):
for batch_idx, (x, y) in enumerate(train_loader):
# x: [b, 1, 28, 28], y:[256]
# [b, 1, 28, 28] ==> [b, 784]
x = x.view(x.size(0), 28*28)
# ==> [b, 10]
out = net(x)
y_onehot = one_hot(y)
# loss = mse(out, y_onehot)
loss = F.mse_loss(out, y_onehot)
optimizer.zero_grad()
loss.backward()
# w' = w - lr*grad
optimizer.step()
train_loss.append(loss.item())
if batch_idx % 10 == 0:
print(epoch, batch_idx, loss.item())
#get optimal w1 b1 w2...
对loss的变化进行可视化,我们可以很清楚的看到loss不断下降
plot_curve(train_loss)
最后我们使用训练好的神经网络对测试集进行预测,准确率有90%多,效果还可以。
total_correct = 0
for x, y in test_loader:
x = x.view(x.size(0), 28*28)
out = net(x)
# out:[b, 10] ==> pred[b]
pred = out.argmax(dim = 1)
correct = pred.eq(y).sum().float().item()
total_correct += correct
total_sum = len(test_loader.dataset)
acc = total_correct / total_sum
print("test accuracy:", acc)
test accuracy: 0.9141
以上就是使用简单的神经网络对手写数字进行辨识,
代码内容出自网易云课程,本篇blog是对该课程进行简单记录和总结