pytorch简单手写数字识别

简单回归问题

只使用numpy编写，代码里有注释

import numpy as np


# y = wx + b
# 计算均方误差
# 𝒍𝒐𝒔𝒔 = 𝑾𝑿 + 𝒃 − 𝒚 
def compute_error_for_line_given_points(b, w, points):
    totalError = 0
    for i in range(0, len(points)):
        x = points[i, 0]
        y = points[i, 1]
        totalError += (y - (w * x + b)) ** 2
    return totalError / float(len(points))


# 计算截距和斜率
# 𝒘′ = 𝒘 − 𝒍𝒓 ∗(𝛁𝒍𝒐𝒔𝒔/𝛁𝒘)
# 𝒍𝒐𝒔𝒔 = 𝑾𝑿 + 𝒃 − 𝒚 
def step_gradient(b_current, w_current, points, learningRate):
    b_gradient = 0
    w_gradient = 0
    N = float(len(points))
    for i in range(0, len(points)):
        x = points[i, 0]
        y = points[i, 1]
        b_gradient += -(2 / N) * (y - ((w_current * x) + b_current))
        w_gradient += -(2 / N) * x * (y - ((w_current * x) + b_current))
    new_b = b_current - (learningRate * b_gradient)
    new_w = w_current - (learningRate * w_gradient)
    return [new_b, new_w]


# 迭代更新截距和斜率
def gradient_descent_runner(points, starting_b, starting_w, learning_rate, num_iterations):
    b = starting_b
    w = starting_w
    for i in range(num_iterations):
        b, w = step_gradient(b, w, np.array(points), learning_rate)
    return [b, w]


def run():
    points = np.genfromtxt("data.csv", delimiter=",")
    learning_rate = 0.0001
    initial_b = 0  # 初始化y截距
    initial_w = 0  # 初始化斜率
    num_iterations = 1000  # 迭代1000次
    print("Starting gradient descent at b = {0}, w = {1}, error = {2}"
          .format(initial_b, initial_w,
                  compute_error_for_line_given_points(initial_b, initial_w, points))
          )
    print("Running...")
    [b, w] = gradient_descent_runner(points, initial_b, initial_w, learning_rate, num_iterations)
    print("After {0} iterations b = {1}, w = {2}, error = {3}".
          format(num_iterations, b, w,
                 compute_error_for_line_given_points(b, w, points))
          )


if __name__ == '__main__':
    run()

结果：

Starting gradient descent at b = 0, w = 0, error = 5565.107834483211
Running...
After 1000 iterations b = 0.08893651993741346, w = 1.4777440851894448, error = 112.61481011613473

手写数字识别

代码，有注释

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 = 512

# step1. 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,))
                               ])),
    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)

# 在装载完成后，我们可以选取其中一个批次的数据进行预览
x, y = next(iter(train_loader))
print(x.shape, y.shape, x.min(), x.max())
plot_image(x, y, 'image sample')


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+b1)
        x = F.relu(self.fc1(x))
        # h2 = relu(h1w2+b2)
        x = F.relu(self.fc2(x))
        # h3 = h2w3+b3
        x = self.fc3(x)

        return x


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, 784]
        x = x.view(x.size(0), 28 * 28)  # 将前面多维度的tensor展平成一维,x.size(0)=b
        # => [b, 10]
        out = net(x)
        # [b, 10]
        y_onehot = one_hot(y)
        # loss = mse(out, y_onehot)
        loss = F.mse_loss(out, y_onehot)
        # 先给梯度清0
        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())

plot_curve(train_loss)
# we get optimal [w1, b1, w2, b2, w3, b3]


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('test acc:', acc)

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')

结果：

torch.Size([512, 1, 28, 28]) torch.Size([512]) tensor(-0.4242) tensor(2.8215)
0 0 0.10631299018859863
0 10 0.09327627718448639
0 20 0.08357524871826172
0 30 0.07842864841222763
0 40 0.07161600142717361
0 50 0.06780468672513962
0 60 0.06343945860862732
0 70 0.061706673353910446
0 80 0.056804824620485306
0 90 0.05541609972715378
0 100 0.052139125764369965
0 110 0.05142538994550705
1 0 0.05038641765713692
1 10 0.04735875129699707
1 20 0.04581252112984657
1 30 0.046893153339624405
1 40 0.04484379291534424
1 50 0.045751187950372696
1 60 0.044139545410871506
1 70 0.04007252678275108
1 80 0.04167523235082626
1 90 0.041462428867816925
1 100 0.039175406098365784
1 110 0.0402270182967186
2 0 0.04034041985869408
2 10 0.03800373896956444
2 20 0.03578241541981697
2 30 0.03599751740694046
2 40 0.03725225850939751
2 50 0.034895461052656174
2 60 0.03328505530953407
2 70 0.033005014061927795
2 80 0.036702465265989304
2 90 0.03656462952494621
2 100 0.03397378325462341
2 110 0.030807316303253174
test acc: 0.8842

精度不高，这简单的使用梯度下降，而且只有3层的神经网络，肯定不太理想，最后可以用CNN卷积神经网络模型，后期测试精度可以达到99.7%