2025.7.17 乱数まとめ【numpy】【torch】
numpyの乱数
numpy.random.rand ... 0-1 の一様乱数
numpy.random.randn ... 正規分布
code:np.py
import numpy as np
print('# rand(3)')
print(np.random.rand(3))
print('# rand(2, 3)')
print(np.random.rand(2, 3))
print('# randn(3)')
print(np.random.randn(3))
print('# randn(2, 3, 4)')
print(np.random.randn(2, 3, 4))
'''
# (3)
0.52418371 0.67171033 0.29514347
# (2, 3)
[0.58346378 0.90481285 0.03869463
0.46989923 0.92635677 0.44764052]
# (3)
-0.93975875 0.10415805 -0.90780392
# (2, 3, 4)
[[-0.79616812 0.30297299 0.25162722 -1.0394509
-0.08196602 -1.70583805 -0.50713506 0.08436172
0.63702657 -0.05040121 -1.6571716 -0.67984694]
[ 0.81127833 -0.67470308 0.19001316 0.78073626
0.29235103 0.96188355 -0.3767867 -0.59544499
0.78829879 0.50815301 0.37267026 0.45669552]]
'''
torchの乱数
torch.rand ... 一様乱数
torch.randn ... 正規分布
どちらも第1引数のタプル・リストで乱数行列の形状を指定する。
code:p.py
import torch as pt
print(pt.rand(1))
print(pt.rand(2,3))
print(pt.randn(2,3,4))
'''
tensor(0.5081)
tensor([0.9388, 0.7614, 0.0600,
0.2595, 0.9622, 0.4446])
tensor([[ 0.6505, 0.8769, 1.8452, 0.3390,
0.3211, -0.9599, 0.9466, -0.1600,
0.7583, -0.7011, 0.2284, -0.4718],
[ 0.3524, 0.8254, 1.6622, 1.2324,
-0.7284, 0.7441, 1.0744, 1.3678,
2.0208, -0.0838, -1.0946, 0.9259]])
'''