切换git用户重新进行项目首次归档

FC_ML_Linear/FC_ML_Least_Squares.py

@@ -0,0 +1,9 @@
+import torch
+
+# 构建超定方程组 Ax=b (3方程2未知数)
+A = torch.tensor([[2.0, 3], [1, 4], [3, 1]])  # 3x2矩阵
+b = torch.tensor([5.0, 6, 4]).reshape(-1,1)    # 3x1向量
+
+# 解法1：正规方程 (A^T A)^-1 A^T b
+solution = torch.linalg.lstsq(A, b).solution    # PyTorch内置最小二乘
+print(f"最小二乘解:\n{solution.numpy()}")

FC_ML_Linear/FC_ML_Linear.py

@@ -0,0 +1,41 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import numpy as np
+import matplotlib.pyplot as plt
+
+# 生成数据
+np.random.seed(0)
+x = np.random.rand(100, 1) * 10  # 生成100个在0到10之间的随机数
+y = 3.5 * x + 2 + np.random.randn(100, 1) * 2  # 真实的线性关系加上一些噪声
+
+# 将numpy数组转换为torch张量
+x_tensor = torch.from_numpy(x).float()
+y_tensor = torch.from_numpy(y).float()
+
+
+model = nn.Linear(in_features=1, out_features=1, bias=True)
+
+criterion = nn.MSELoss()
+optimizer = optim.SGD(model.parameters(), lr=0.01)
+
+epochs = 100
+for epoch in range(epochs):
+    # 前向传播
+    outputs = model(x_tensor)
+    loss = criterion(outputs, y_tensor)
+
+    # 反向传播和优化
+    optimizer.zero_grad()
+    loss.backward()
+    optimizer.step()
+
+    if (epoch + 1) % 10 == 0:
+        print(f'Epoch [{epoch + 1}/{epochs}], Loss: {loss.item():.4f}')
+
+# 可视化原始数据和拟合结果
+predicted = model(x_tensor).detach().numpy()  # detach()用于从计算图中移除张量，避免未来的梯度计算干扰。
+plt.scatter(x, y, label='Original data')
+plt.plot(x, predicted, color='red', label='Fitted line')
+plt.legend()
+plt.show()