torch.normal()、torch.randn()和torch.rand()简要介绍

定义

torch.normal(means,std,size)

torch.randn(size)

torch.rand(size)

用法

• torch.normal()：返回从均值means和标准差std的离散正态分布中抽取随机张量，
• torch.randn()：生成满足标准正态分布（0~1）的随机张量
• torch.rand()：返回从区间[0, 1)的均匀分布中抽取的一组随机数

示例

import torch
import numpy as np

a = torch.normal(10,2,(2,3))
b = torch.randn((3,4))
c = torch.rand((4,5))
print(a)
print(b)
print(c)

>>>tensor([[13.6814,  6.1459,  8.2079],
           [ 8.4074,  7.3686, 12.5890]])

tensor([[ 0.7213, -0.6970,  1.5204,  0.0059],
        [ 2.3095, -0.6058,  0.9244,  0.4169],
        [-0.0146,  0.7050,  0.4710, -0.2338]])

tensor([[0.4213, 0.6088, 0.3365, 0.6926, 0.4170],
        [0.0060, 0.6874, 0.2774, 0.5824, 0.9821],
        [0.9799, 0.6080, 0.8078, 0.4144, 0.4762],
        [0.3729, 0.8740, 0.8804, 0.8167, 0.1810]])
​