DEV Community

Super Kai (Kazuya Ito)
Super Kai (Kazuya Ito)

Posted on • Edited on

rand and rand_like in PyTorch

Buy Me a Coffee

*Memos:

rand() can create the 0D or more D tensor of the zero or more random floating-point numbers(Default) or complex numbers between 0 and 1(0<=x<1) as shown below:

*Memos:

  • rand() can be used with torch but not with a tensor.
  • The 1st or more arguments with torch are size(Required-Type:int, tuple of int, list of int or size()).
  • There is dtype argument with torch(Optional-Default:None-Type:dtype): *Memos:
  • There is device argument with torch(Optional-Default:None-Type:str, int or device()): *Memos:
  • There is requires_grad argument with torch(Optional-Default:False-Type:bool): *Memos:
    • requires_grad= must be used.
    • My post explains requires_grad argument.
  • There is out argument with torch(Optional-Default:None-Type:tensor): *Memos:
    • out= must be used.
    • My post explains out argument.
import torch

torch.rand(size=())
torch.rand(size=torch.tensor(8).size())
# tensor(0.2965)

torch.rand(size=(0,))
torch.rand(0)
torch.rand(size=torch.tensor([]).size())
# tensor([])

torch.rand(size=(3,))
torch.rand(3)
torch.rand(size=torch.tensor([8, 3, 6]).size())
# tensor([0.5138, 0.6443, 0.8991])

torch.rand(size=(3, 2))
torch.rand(3, 2)
torch.rand(size=torch.tensor([[8, 3], [6, 0], [2, 9]]).size())
# tensor([[0.0141, 0.5785],
#         [0.1218, 0.9181],
#         [0.6805, 0.2000]])

torch.rand(size=(3, 2, 4))
torch.rand(3, 2, 4)
# tensor([[[0.1121, 0.0019, 0.8928, 0.8047],
#          [0.2547, 0.4683, 0.1682, 0.1713]],
#         [[0.5103, 0.1068, 0.5911, 0.7893],
#          [0.0854, 0.3679, 0.4761, 0.6386]],
#         [[0.4480, 0.0795, 0.3199, 0.0032],
#          [0.4238, 0.3069, 0.9529, 0.4462]]])

torch.rand(size=(3, 2, 4), dtype=torch.complex64)
torch.rand(3, 2, 4, dtype=torch.complex64)
# tensor([[[0.7756+0.0131j, 0.4101+0.8469j, 0.7885+0.6149j, 0.1891+0.7311j],
#          [0.9582+0.4655j, 0.7719+0.1614j, 0.6985+0.5411j, 0.9334+0.7705j]],
#         [[0.9722+0.4131j, 0.4887+0.4382j, 0.4342+0.0959j, 0.1571+0.3554j],
#          [0.1554+0.0313j, 0.2604+0.5255j, 0.0083+0.6250j, 0.9909+0.3338j]],
#         [[0.2127+0.8406j, 0.5262+0.7138j, 0.1354+0.1573j, 0.3121+0.0703j],
#          [0.3213+0.3891j, 0.6951+0.5847j, 0.0853+0.5330j, 0.3310+0.6671j]]])
Enter fullscreen mode Exit fullscreen mode

rand_like() can replace the zero or more floating-point numbers or complex numbers of a 0D or more D tensor with the zero or more random floating-point numbers or complex numbers between 0 and 1(0<=x<1) as shown below:

*Memos:

  • rand_like() can be used with torch but not with a tensor.
  • The 1st argument with torch is input(Required-Type:tensor of float or complex).
  • There is dtype argument with torch(Optional-Default:None-Type:dtype): *Memos:
    • If it's None, it's inferred from input.
    • dtype= must be used.
    • My post explains dtype argument.
  • There is device argument with torch(Optional-Default:None-Type:str, int or device()): *Memos:
    • If it's None, it's inferred from input.
    • device= must be used.
    • My post explains device argument.
  • There is requires_grad argument with torch(Optional-Default:False-Type:bool): *Memos:
    • requires_grad= must be used.
    • My post explains requires_grad argument.
import torch

my_tensor = torch.tensor(7.)

torch.rand_like(input=my_tensor)
# tensor(0.3634)

my_tensor = torch.tensor([7., 4., 5.])

torch.rand_like(input=my_tensor)
# tensor([0.9000, 0.1314, 0.5240])

my_tensor = torch.tensor([[7., 4., 5.], [2., 8., 3.]])

torch.rand_like(input=my_tensor)
# tensor([[0.7508, 0.5857, 0.1658],
#         [0.3327, 0.1836, 0.3776]])

my_tensor = torch.tensor([[[7., 4., 5.], [2., 8., 3.]], 
                          [[6., 0., 1.], [5., 9., 4.]]])
torch.rand_like(input=my_tensor)
# tensor([[[0.8467, 0.3134, 0.9486], [0.3134, 0.1809, 0.1740]],
#         [[0.2624, 0.3929, 0.2138], [0.2034, 0.9147, 0.3421]]])

my_tensor = torch.tensor([[[7.+4.j, 4.+2.j, 5.+3.j],
                           [2.+5.j, 8.+1.j, 3.+9.j]],
                          [[6.+9.j, 0.+3.j, 1.+8.j],
                           [5.+3.j, 9.+4.j, 4.+6.j]]])
torch.rand_like(input=my_tensor)
# tensor([[[0.9179+0.7615j, 0.6811+0.3994j, 0.6854+0.3131j],
#          [0.7352+0.0108j, 0.0999+0.4509j, 0.9586+0.1530j]],
#         [[0.5253+0.7685j, 0.5731+0.0358j, 0.9499+0.8943j],
#          [0.7742+0.3421j, 0.0419+0.8636j, 0.1014+0.8507j]]])
Enter fullscreen mode Exit fullscreen mode

Image of Timescale

Timescale – the developer's data platform for modern apps, built on PostgreSQL

Timescale Cloud is PostgreSQL optimized for speed, scale, and performance. Over 3 million IoT, AI, crypto, and dev tool apps are powered by Timescale. Try it free today! No credit card required.

Try free

Top comments (0)

Billboard image

The Next Generation Developer Platform

Coherence is the first Platform-as-a-Service you can control. Unlike "black-box" platforms that are opinionated about the infra you can deploy, Coherence is powered by CNC, the open-source IaC framework, which offers limitless customization.

Learn more