菜鸟文档

基础代码

import numpy as np
import torch
import torch.nn as nn
import torch.optim as optim

# 创建一个简单的模型
class Model(nn.Module):
    def __init__(self):
        super(Model, self).__init__()
        self.linear = nn.Linear(10, 1)  # 输入特征维度为3，输出为1
        self.sofamax = nn.Softmax(dim=1) # 输入特征维度为3，输出为1

        self.linear1 = nn.Linear(10, 1)  # 输入特征维度为3，输出为1
        # self.sofamax1 = nn.Softmax(dim=1) # 输入特征维度为3，输出为1
        self.sig = nn.Sigmoid()

    def forward(self, x):
        out = self.linear(x)
        out = self.sofamax(out)
        out = self.linear1(x)
        # out = self.sofamax1(out)
        out = self.sig(out)
        return out

# 准备数据
# x_train = torch.tensor([[1, 2, 3, 1, 2, 1, 2, 3, 1, 2], [4, 5, 6,5, 6, 4, 5, 6,5, 6], [4, 5, 6,5, 6, 7, 8, 7, 8, 9]], dtype=torch.float32)
# y_train = torch.tensor([[1], [1], [0]], dtype=torch.float32)

x_train = torch.tensor(np.random.random((100, 10)), dtype=torch.float32)
y_train = torch.tensor(np.random.randint(0, 2, (100, 1)), dtype=torch.float32)

print("x_train shape:", x_train.shape)
print("y_train shape:", y_train.shape)

# 创建模型实例、损失函数和优化器
model = Model()
criterion = nn.BCELoss()
optimizer = optim.SGD(model.parameters(), lr=0.01)

# 训练模型
num_epochs = 1000
for epoch in range(num_epochs):
    # 前向传播
    outputs = model(x_train)
    loss = criterion(outputs, y_train)

    # 反向传播和优化
    optimizer.zero_grad()  # 清零梯度
    loss.backward()  # 反向传播
    optimizer.step()  # 更新权重

    if (epoch+1) % 100 == 0:
        print(f"Epoch [{epoch+1}/{num_epochs}], Loss: {loss.item():.4f}")

# 使用训练好的模型进行预测
x_test = torch.tensor([[1, 4, 6, 2, 1, 6, 2, 4, 6, 1], [3, 6, 9, 2, 4, 6, 2, 4, 6, 2]], dtype=torch.float32)
preds = model(x_test)
print("预测结果：")
print(preds)

x_train shape: torch.Size([100, 10]) y_train shape: torch.Size([100, 1]) Epoch [100/1000], Loss: 0.6853 Epoch [200/1000], Loss: 0.6840 Epoch [300/1000], Loss: 0.6831 Epoch [400/1000], Loss: 0.6823 Epoch [500/1000], Loss: 0.6816 Epoch [600/1000], Loss: 0.6809 Epoch [700/1000], Loss: 0.6802 Epoch [800/1000], Loss: 0.6796 Epoch [900/1000], Loss: 0.6790 Epoch [1000/1000], Loss: 0.6784 预测结果： tensor([[0.9321], [0.9668]], grad_fn=)

机器学习记录

基础代码