Numpy random number
- [0,1) evenly distributed number of randn
r1 = np.random.rand()
r2 = np.random.rand(2,5)
r1:
0.6063522922681436
r2:
[[0.77765149 0.02672804 0.59697894 0.90333184 0.23724221]
[0.74653515 0.25098751 0.30388215 0.60496512 0.92437331]]
- Mean 0, variance 1 rand and use the same
r3 = np.random.randn()
r4 = np.random.randn(2,5)
# 对上边的升级 产生指定的均值和方差
r5 = 2.5 * np.random.randn(2,5) + 3 # 2.5是标准差,3是均值
r3:
1.2866481303331248
r4:
[[-0.69264271 0.14013687 0.65103358 -0.24985809 0.64292644]
[-0.13125458 0.73133279 0.27202034 -1.85005286 -0.42510794]]
r5:
[[ 3.60175624 -0.94849954 5.56319813 -1.13279464 5.5410862 ]
[ 4.1107094 3.32757096 2.81362529 5.85241203 0.04886216]]
- Randomly generating a [0,1) real number
r6 = np.random.random()
r7 = 5 * np.random.random() - 5
r6:
0.9805701484882132
r7:
-2.2129076587989682
r8 = np.random.randint(10)
r9 = np.random.randint(low=5,high=10,size=(2,4))
r8:
8
r9:
[[5 6 5 8]
[9 8 8 9]]
- Usage substantially uniform from the specified range to generate a random float within
r10 = np.random.uniform()
r11 = np.random.uniform(low=[1,2],high=[2,3],size=(2,2))
r10:
0.07179463907175299
r11:
[[1.88509028 2.05329813]
[1.60257834 2.72071315]]
r12 = np.random.choice(['正','反'],size=2)
r12:
['正' '正']
- In the original basis scrambled on
r13 = np.arange(1,5)
np.random.shuffle(r13)
r13:
[2 1 4 3]
Pytorch random number
- Randomly generate [0-1) random numbers
r14 = torch.rand(2,4)
r14:
tensor([[0.3398, 0.8372, 0.9900, 0.2805],
[0.5386, 0.9997, 0.1365, 0.1415]])
- The standard normal distribution too mean 0, variance 1
r15 = torch.randn(2,4)
r15:
tensor([[-0.8925, -0.0280, -0.7318, 0.3898],
[-0.6744, 0.3835, 1.3782, 0.1880]])
r16 = torch.linspace(1,10,10)
r17 = np.linspace(1,10,10)
r16:
tensor([ 1., 2., 3., 4., 5., 6., 7., 8., 9., 10.])
r17:
[ 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.]