update torch4ms tests #2234

tests/torch4ms/test_simple_op.py

@@ -0,0 +1,59 @@
+import torch4ms 
+import torch 
+import mindspore 
+import numpy as np
+
+env = torch4ms.default_env()
+env.__enter__() 
+
+def test_matrix_operations():
+    """测试矩阵乘法和加法组合运算"""
+    # 创建测试数据
+    np_x = np.array([[1.0, 2.0], [3.0, 4.0]], dtype=np.float32)
+    np_y = np.array([[5.0, 6.0], [7.0, 8.0]], dtype=np.float32)
+    np_z = np.array([[9.0, 10.0], [11.0, 12.0]], dtype=np.float32)
+
+    x = torch.tensor(np_x)
+    y = torch.tensor(np_y)
+    z = torch.tensor(np_z)
+    result = torch.matmul(x, y) + z
+
+    print(f"x = {x}")
+    print(f"y = {y}")
+    print(f"z = {z}")
+    print(f"x * y + z = {result}")
+
+    expected = np.matmul(np_x, np_y) + np_z
+    print(f"\n预期结果:")
+    print(f"{expected}")
+
+    np_result = result.detach().numpy()
+    print(f"\n数值验证:")
+    print(f"结果是否接近预期: {np.allclose(np_result, expected, atol=1e-5)}")
+
+def test_activation_functions():
+    """测试激活函数"""
+    np_data = np.array([[-1.0, 0.0, 1.0], [-0.5, 0.5, 1.5]], dtype=np.float32)
+    x = torch.tensor(np_data)
+
+    print("\n" + "="*40)
+    print("测试激活函数:")
+    print(f"输入: {x}")
+
+    relu_result = torch.relu(x)
+    print(f"\nReLU结果: {relu_result}")
+
+    sigmoid_result = torch.sigmoid(x)
+    print(f"Sigmoid结果: {sigmoid_result}")
+
+    tanh_result = torch.tanh(x)
+    print(f"Tanh结果: {tanh_result}")
+
+if __name__ == "__main__":
+    print("PyTorch版本: {}".format(torch.__version__))
+    print("MindSpore版本: {}".format(mindspore.__version__))
+    print("="*40)
+
+    test_matrix_operations()
+    test_activation_functions()
+    env.__exit__(None, None, None)