From 71642c690d512c2d4b0e25a36494e1b11d5469f0 Mon Sep 17 00:00:00 2001
From: root
 <root@inss-62a9d72c3b452e2a31780ace-iprj68564c64e46d2454ac0aba12.inss-62a9d72c3b452e2a31780ace-iprj68564c64e46d2454ac0aba12.inscode.svc.cluster.local>
Date: Wed, 23 Jul 2025 14:57:00 +0800
Subject: [PATCH] Wed Jul 23 14:57:00 CST 2025 inscode

---
 test02.py           | 31 ----------------
 test02.py/test03.py | 89 +++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 89 insertions(+), 31 deletions(-)
 delete mode 100644 test02.py
 create mode 100644 test02.py/test03.py

diff --git a/test02.py b/test02.py
deleted file mode 100644
index 93bbc26..0000000
--- a/test02.py
+++ /dev/null
@@ -1,31 +0,0 @@
-import torch
-
-# 创建一个 2x3 的全 0 张量
-a = torch.zeros(2, 3)
-print(a)
-
-# 创建一个 2x3 的全 1 张量
-b = torch.ones(2, 3)
-print(b)
-
-# 创建一个 2x3 的随机数张量
-c = torch.randn(2, 3)
-print(c)
-
-# 从 NumPy 数组创建张量
-import numpy as np
-numpy_array = np.array([[1, 2], [3, 4]])
-tensor_from_numpy = torch.from_numpy(numpy_array)
-print(tensor_from_numpy)
-
-# 在指定设备（CPU/GPU）上创建张量
-device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
-d = torch.randn(2, 3, device=device)
-print(torch.cuda.is_available())
-print(d)
-
-# 逐元素乘法（要求形状相同）
-X = torch.tensor([[1, 2], [3, 4]])
-Y = torch.tensor([[5, 6], [7, 8]])
-Z = X * Y  # 或 torch.mul(X, Y)
-print("逐元素乘法结果:\n", Z)
\ No newline at end of file
diff --git a/test02.py/test03.py b/test02.py/test03.py
new file mode 100644
index 0000000..c4d1edf
--- /dev/null
+++ b/test02.py/test03.py
@@ -0,0 +1,89 @@
+import torch
+import torch.nn as nn
+import torch.optim as optim
+import numpy as np
+import matplotlib.pyplot as plt
+
+# 设置随机种子保证可重复性
+torch.manual_seed(42)
+
+# ==========================
+# 1. 生成正弦函数数据
+# ==========================
+def generate_sine_data(seq_length=50, total_points=1000):
+    x = np.linspace(0, 100, total_points)
+    y = np.sin(x)
+    
+    # 构造输入序列和目标值
+    X, Y = [], []
+    for i in range(len(y) - seq_length):
+        X.append(y[i:i+seq_length])  # 输入序列
+        Y.append(y[i+seq_length])    # 下一个值（目标）
+    
+    return np.array(X), np.array(Y)
+
+seq_length = 50
+X, y = generate_sine_data(seq_length=seq_length)
+
+# 转换为 PyTorch 张量
+X = torch.FloatTensor(X).unsqueeze(-1)  # 形状: (样本数, seq_length, 1)
+y = torch.FloatTensor(y).unsqueeze(-1)  # 形状: (样本数, 1)
+
+# ==========================
+# 2. 定义 RNN 模型
+# ==========================
+class SimpleRNN(nn.Module):
+    def __init__(self, input_size=1, hidden_size=32, output_size=1):
+        super().__init__()
+        self.rnn = nn.RNN(input_size, hidden_size, batch_first=True)
+        self.fc = nn.Linear(hidden_size, output_size)
+    
+    def forward(self, x):
+        # x 形状: (batch_size, seq_length, input_size)
+        out, _ = self.rnn(x)  # out 形状: (batch_size, seq_length, hidden_size)
+        out = self.fc(out[:, -1, :])  # 取最后一个时间步的输出
+        return out
+
+model = SimpleRNN()
+criterion = nn.MSELoss()
+optimizer = optim.Adam(model.parameters(), lr=0.01)
+
+# ==========================
+# 3. 训练模型
+# ==========================
+epochs = 200
+losses = []
+
+for epoch in range(epochs):
+    optimizer.zero_grad()
+    outputs = model(X)
+    loss = criterion(outputs, y)
+    loss.backward()
+    optimizer.step()
+    
+    losses.append(loss.item())
+    if (epoch+1) % 20 == 0:
+        print(f'Epoch [{epoch+1}/{epochs}], Loss: {loss.item():.4f}')
+
+# 绘制损失曲线
+plt.figure(figsize=(12, 4))
+plt.plot(losses)
+plt.xlabel('Epoch')
+plt.ylabel('Loss')
+plt.title('Training Loss')
+plt.show()
+
+# ==========================
+# 4. 预测并可视化结果
+# ==========================
+model.eval()
+with torch.no_grad():
+    predictions = model(X)
+
+# 可视化前 200 个点的预测结果
+plt.figure(figsize=(12, 6))
+plt.plot(y[:200].numpy(), label='True')
+plt.plot(predictions[:200].numpy(), label='Predicted')
+plt.legend()
+plt.title('RNN Sine Wave Prediction')
+plt.show()
\ No newline at end of file
-- 
GitLab