PyTorch Create Tensor with Random Values and Specific Shape

Create PyTorch Tensor with Ramdom Values

To create a random tensor with specific shape, use torch.rand() function with shape passed as argument to the function. torch.rand() function returns tensor with random values generated in the specified shape.

Example

Following is a simple example, where in we created a tensor of specific size filled with random values.

import torch
#create tensor with random data
rand_tensor = torch.rand((2, 5))
#print tensor
print(rand_tensor)
Run
PyTorch create tensor with random values and specified size

Create PyTorch Tensor with Random Values less than a Specific Maximum Value

By default, pytorch.rand() function generates tensor with floating point values ranging between 0 and 1.

Example

In the following example, we will create a tensor with random values that are less than 8.

import torch
#create tensor with random data, and multiply with a scalar
rand_tensor = 8*torch.rand((2, 5))
#print tensor
print(rand_tensor)
Run

Create Tensor with Random Values from a Range

You can also create a PyTorch Tensor with random values belonging to a specific range (min, max).

Example

In the following example, we have taken a range of (4, 8) and created a tensor, with random values being picked from the range (4, 8).

import torch
#range
max = 8
min = 4
#create tensor with random values in range (min, max)
rand_tensor = (max-min)*torch.rand((2, 5)) + min
#print tensor
print(rand_tensor)
Run