深度学习作业4-BP算法

1、numpy程序

import numpy as np

# 初始权重
w1, w2, w3, w4, w5, w6, w7, w8 = 0.2, -0.4, 0.5, 0.6, 0.1, -0.5, -0.3, 0.8
# 输入和目标输出
x1, x2 = 0.5, 0.3
y1, y2 = 0.23, -0.07
print("输入值 x1, x2:", x1, x2)
print("输出值 y1, y2:", y1, y2)

# Sigmoid 激活函数
def sigmoid(z):
    return 1 / (1 + np.exp(-z))

# 前向传播
def forward_propagate(x1, x2, y1, y2, w1, w2, w3, w4, w5, w6, w7, w8):
    in_h1 = w1 * x1 + w3 * x2
    out_h1 = sigmoid(in_h1)
    in_h2 = w2 * x1 + w4 * x2
    out_h2 = sigmoid(in_h2)

    in_o1 = w5 * out_h1 + w7 * out_h2
    out_o1 = sigmoid(in_o1)
    in_o2 = w6 * out_h1 + w8 * out_h2
    out_o2 = sigmoid(in_o2)

    error = (1 / 2) * (out_o1 - y1) ** 2 + (1 / 2) * (out_o2 - y2) ** 2
    return out_o1, out_o2, out_h1, out_h2, error

# 反向传播
def back_propagate(out_o1, out_o2, out_h1, out_h2, y1, y2, w5, w7, w6, w8):
    d_o1 = out_o1 - y1
    d_o2 = out_o2 - y2

    d_w5 = d_o1 * out_o1 * (1 - out_o1) * out_h1
    d_w7 = d_o1 * out_o1 * (1 - out_o1) * out_h2
    d_w6 = d_o2 * out_o2 * (1 - out_o2) * out_h1
    d_w8 = d_o2 * out_o2 * (1 - out_o2) * out_h2

    d_w1 = (d_o1 * out_o1 * (1 - out_o1) * w5 + d_o2 * out_o2 * (1 - out_o2) * w6) * out_h1 * (1 - out_h1) * x1
    d_w3 = (d_o1 * out_o1 * (1 - out_o1) * w5 + d_o2 * out_o2 * (1 - out_o2) * w6) * out_h1 * (1 - out_h1) * x2
    d_w2 = (d_o1 * out_o1 * (1 - out_o1) * w7 + d_o2 * out_o2 * (1 - out_o2) * w8) * out_h2 * (1 - out_h2) * x1
    d_w4 = (d_o1 * out_o1 * (1 - out_o1) * w7 + d_o2 * out_o2 * (1 - out_o2) * w8) * out_h2 * (1 - out_h2) * x2

    return d_w1, d_w2, d_w3, d_w4, d_w5, d_w6, d_w7, d_w8

# 更新权值
def update_w(w1, w2, w3, w4, w5, w6, w7, w8, d_w1, d_w2, d_w3, d_w4, d_w5, d_w6, d_w7, d_w8, step):
    w1 = w1 - step * d_w1
    w2 = w2 - step * d_w2
    w3 = w3 - step * d_w3
    w4 = w4 - step * d_w4
    w5 = w5 - step * d_w5
    w6 = w6 - step * d_w6
    w7 = w7 - step * d_w7
    w8 = w8 - step * d_w8
    return w1, w2, w3, w4, w5, w6, w7, w8

# 主程序
if __name__ == "__main__":
    step = 1  # 学习率
    print("权值w1-w8初始值:", round(w1, 2), round(w2, 2), round(w3, 2), round(w4, 2), round(w5, 2), round(w6, 2), round(w7, 2), round(w8, 2))

    for i in range(1000):
        out_o1, out_o2, out_h1, out_h2, error = forward_propagate(x1, x2, y1, y2, w1, w2, w3, w4, w5, w6, w7, w8)
        d_w1, d_w2, d_w3, d_w4, d_w5, d_w6, d_w7, d_w8 = back_propagate(out_o1, out_o2, out_h1, out_h2, y1, y2, w5, w7, w6, w8)
        w1, w2, w3, w4, w5, w6, w7, w8 = update_w(w1, w2, w3, w4, w5, w6, w7, w8, d_w1, d_w2, d_w3, d_w4, d_w5, d_w6, d_w7, d_w8, step)

        # 在第1轮，第5轮，第50轮，第1000轮时输出结果
        if i + 1 in [1, 5, 50, 1000]:
            print(f"\n=====第 {i+1} 轮=====")
            print(f"隐藏层输出: h1={round(out_h1, 5)}, h2={round(out_h2, 5)}")
            print(f"预测值: o1={round(out_o1, 5)}, o2={round(out_o2, 5)}")
            print(f"损失函数(均方误差): {round(error, 5)}")
            print(f"更新后的权值: w1={round(w1, 2)}, w2={round(w2, 2)}, w3={round(w3, 2)}, w4={round(w4, 2)}, w5={round(w5, 2)}, w6={round(w6, 2)}, w7={round(w7, 2)}, w8={round(w8, 2)}")

2、PyTorch程序

import torch

# 输入和目标输出
x = [0.5, 0.3]  # x0, x1
y = [0.23, -0.07]  # y0, y1
print("输入值 x0, x1:", x[0], x[1])
print("输出值 y0, y1:", y[0], y[1])

# 初始化权重
w = [torch.Tensor([0.0]), torch.Tensor([0.0]), torch.Tensor([0.0]), torch.Tensor([0.0]),
     torch.Tensor([0.0]), torch.Tensor([0.0]), torch.Tensor([0.0]), torch.Tensor([0.0])]
for i in range(0, 8):
    w[i].requires_grad = True

# 打印初始权重
print("权值w0-w7:")
for i in range(0, 8):
    print(w[i].data, end="  ")

# 前向传播函数
def forward_propagate(x):
    in_h1 = w[0] * x[0] + w[2] * x[1]
    out_h1 = torch.sigmoid(in_h1)
    in_h2 = w[1] * x[0] + w[3] * x[1]
    out_h2 = torch.sigmoid(in_h2)

    in_o1 = w[4] * out_h1 + w[6] * out_h2
    out_o1 = torch.sigmoid(in_o1)
    in_o2 = w[5] * out_h1 + w[7] * out_h2
    out_o2 = torch.sigmoid(in_o2)

    return out_o1, out_o2

# 损失函数
def loss(x, y):
    y_pre = forward_propagate(x)  # 前向传播
    loss_mse = (1 / 2) * (y_pre[0] - y[0]) ** 2 + (1 / 2) * (y_pre[1] - y[1]) ** 2
    return loss_mse

# 主程序
if __name__ == "__main__":
    step = 10  # 步长
    for k in range(1000):  # 训练1000轮
        l = loss(x, y)  # 前向传播，求 Loss，构建计算图
        l.backward()  # 反向传播，求出计算图中所有梯度存入w中

        # 每轮输出的判断
        if k + 1 in [1, 5, 50, 1000]:  # 在第1, 5, 50, 1000轮时输出结果
            print(f"\n=====第 {k+1} 轮=====")
            print("损失函数(均方误差)：", l.item())
            print("w的梯度: ", end="  ")
            for i in range(0, 8):
                print(round(w[i].grad.item(), 2), end="  ")
            print("\n更新后的权值w:")
            for i in range(0, 8):
                print(w[i].data, end="  ")

        # 更新权值
        for i in range(0, 8):
            w[i].data = w[i].data - step * w[i].grad.data  # 更新权值
            w[i].grad.data.zero_()  # 清零梯度