Torch与Numpy简单比较（5）

 （1）numpy array 和 torch tensor之间相互转换

import  torch
import  numpy as np

np_data = np.arange(6).reshape((2, 3))
torch_data = torch.from_numpy(np_data)
tensor2array = torch_data.numpy()

print("\nnp_data:\n",np_data, "\ntorch_data:\n",torch_data, "\ntensor2array:\n", tensor2array)

（2）numpy与torch数学运算

#绝对值计算
import torch
import numpy as np
data = [-1, -2, 1, 2]
tensor = torch.FloatTensor(data) #转换成32位浮点 tensor

print("\nabs\n",'\nnumpy: \n', np.abs(data), '\ntorch:\n', torch.abs(tensor))

#sin 三角函数 sin
print("\nsin\n",'\nnumpy: \n', np.sin(data),
       '\ntorch: \n', torch.sin(tensor))

#mean 均值
print("\nmean\n",'\nnumpy: \n', np.mean(data),
       '\ntorch: \n', torch.mean(tensor))

（3）numpy与torch矩阵运算

import numpy as np
import torch
data  = [[1,2], [3,4]]
tensor = torch.FloatTensor(data)
print('\nmatrix multiplication\n','\nnumpy: \n',np.matmul(data, data),
    '\ntorch:\n', torch.mm(tensor, tensor))

注：文章内容主要引用：https://www.pytorchtutorial.com/

了解更多关于《计算机视觉与图形学》相关知识，请关注公众号：

下载我们视频中代码和相关讲义，请在公众号回复：计算机视觉课程资料