网络模型
设置三层网络，一般最后一层激活函数不选择relu
任务步骤
手写数字识别任务共有四个步骤：
1、数据加载--Load Data
2、构建网络--Build Model
3、训练--Train
4、测试--Test
实战
1、导入各种需要的包
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 minist_utils import plot_image, plot_curve, one_hot ##自写文件
minist_utils:
2、加载数据
batch_size = 512
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, ))
])),
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, ))
])),
batch_size=batch_size, shuffle=False
取一些样本看数据的shape以及图片内容
x, y = next(iter(train_loader))
print(x.shape, y.shape, x.min(), x.max())
plot_image(x, y, 'image sample')
注：经过load加载处理后的数据集包含x（图像信息）和y（标签信息）
next(iter())的用法是取一组样本，重复运行可以依次顺序取样，直到样本被取完
可在csdn自行搜索学习了解
3、网络构建
按之前设想的三层线性模型嵌套的思想搭建模型，为了模型简单，第三层不加激活函数。
class Net(nn.Module):
def __init__(self):
super(Net, self).__init__()
# xw+b
self.fc1 = nn.Linear(28*28, 256) #输入特征数，输出特征数
self.fc2 = nn.Linear(256, 64) #256，64是根据经验判断
self.fc3 = nn.Linear(64, 10) #最开始的28*28和输出的10是一定的
def forward(self, x):
# x: [b, 1, 28, 28]
# h1 = relu(xw1 + b1)
x = F.relu(self.fc1(x)) #输入x后第一次线性模型得到H1作第二层输入
# h2 = relu(h1w2 + b2)
x = F.relu(self.fc2(x)) #输入H1得到H2作第三层输入
# h3 = h2w3 + b3
x = self.fc3(x) #输入H3得到最终结果，维度为10
return x
4、模型训练
net = Net()
# [w1, b1, w2, b2, w3, b3]
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: [512]
# [b, 1, 28, 28] => [b, feature] 全连接层只能接受这样的数据
x = x.view(x.size(0), 28*28)
# => [b, 10]
out = net(x)
# [b, 10]
y_onehot = one_hot(y)
# loss = mse(out, y_onehot)
loss = F.mse_loss(out, y_onehot)
optimizer.zero_grad()
loss.backward() # 梯度计算过程
# w` = w - lr * grad
optimizer.step() # 优化更新w,b
train_loss.append(loss.item())
if batch_idx % 10 == 0:
print(epoch, batch_idx, loss.item())
plot_curve(train_loss)
5、测试
1、计算准确率acc
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_num = len(test_loader.dataset)
acc = total_correct / total_num
print(("acc:", acc))
2、展示部分测试样本原图以及预测标签结果
x, y =next(iter(test_loader))
out = net(x.view(x.size(0), 28*28))
pred = out.argmax(dim=1)
plot_image(x, pred, 'test')
