【python pytorch】Pytorch 基础知识

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/u013421629/article/details/84778121

包含知识点：

• 张量
• 数学操作
• 数理统计
• 比较操作

#-*-coding:utf-8-*-

import numpy as np
np.set_printoptions(suppress=True)
import torch



# 构造一个4*5 的矩阵
z=torch.Tensor(4,5)
print(z)

# 两个矩阵进行加法操作
y=torch.rand(4,5)

print(z+y)
# 另一种表示
print(torch.add(z,y))


# 将tensor 转换为numpy
b=y.numpy()

print(b)


# 数学操作绝对值
kk=torch.abs(torch.FloatTensor([-4,6,90]))
print(kk)

# 均值(行操作)
print(torch.mean(kk,0))

# 比较操作

m1=torch.equal(torch.Tensor([1,2]),torch.Tensor([1,2]))

m2=torch.equal(torch.Tensor([1,2]),torch.Tensor([2,2]))

m3=torch.eq(torch.Tensor([1,2]),torch.Tensor([2,2]))
m4=torch.gt(torch.Tensor([1,2]),torch.Tensor([2,2]))

print(m1)

print(m2)

print(m3)

print(m4)

运行结果：

tensor([[ 0.0000,  0.0000,  0.0000,  0.0000, -3.7296],
        [ 0.0000, -8.2118,  0.0000,  0.0000,  0.0000],
        [ 0.0000,  0.0000, -4.0750,  0.0000, -8.2119],
        [ 0.0000,  0.0000,  0.0000,  0.0000,  0.0000]])
tensor([[ 0.3490,  0.7795,  0.1428,  0.2517, -3.1552],
        [ 0.0427, -7.5753,  0.1780,  0.7305,  0.7264],
        [ 0.2967,  0.2977, -3.8018,  0.2856, -8.0059],
        [ 0.9123,  0.6403,  0.8935,  0.9008,  0.6926]])
tensor([[ 0.3490,  0.7795,  0.1428,  0.2517, -3.1552],
        [ 0.0427, -7.5753,  0.1780,  0.7305,  0.7264],
        [ 0.2967,  0.2977, -3.8018,  0.2856, -8.0059],
        [ 0.9123,  0.6403,  0.8935,  0.9008,  0.6926]])
[[0.34903067 0.7795371  0.14277744 0.25165677 0.57442063]
 [0.04269707 0.63649714 0.17801785 0.73047435 0.72639245]
 [0.29670775 0.29770297 0.27317053 0.28561223 0.20602047]
 [0.91231096 0.6403226  0.8934667  0.90082955 0.69256335]]
tensor([  4.,   6.,  90.])
tensor(33.3333)
True
False
tensor([ 0,  1], dtype=torch.uint8)
tensor([ 0,  0], dtype=torch.uint8)

Process finished with exit code 0

中文教程：
https://pytorch.apachecn.org/#/