Data Structure
Networking
RDBMS
Operating System
Java
MS Excel
iOS
HTML
CSS
Android
Python
C Programming
C++
C#
MongoDB
MySQL
Javascript
PHP
Physics
Chemistry
Biology
Mathematics
English
Economics
Psychology
Social Studies
Fashion Studies
Legal Studies
- Selected Reading
- UPSC IAS Exams Notes
- Developer's Best Practices
- Questions and Answers
- Effective Resume Writing
- HR Interview Questions
- Computer Glossary
- Who is Who
How to rescale a tensor in the range [0, 1] and sum to 1 in PyTorch?
We can rescale an n-dimensional input Tensor such that the elements lie within the range [0,1] and sum to 1. To do this, we can apply the Softmax() function. We can rescale the n-dimensional input tensor along a particular dimension. The size of the output tensor is the same as the input tensor.
Syntax
torch.nn.Softmax(dim)
Parameters
dim – The dimension along which the Softmax is computed.
Steps
We could use the following steps to crop an image at random location with given size −
Import the required library. In all the following examples, the required Python library is torch. Make sure you have already installed it.
import torch
Define a n-dimensional input tensor input.
input = torch.randn(5,2)
Define the Softmax function passing the dimension dim as an optional parameter.
softmax = torch.nn.Softmax(dim = 1)
Apply the above defined Softmax function on the input tensor input.
output = softmax(input)
Print the tensor containing Softmax values.
print(output)
Example 1
The following Python program rescales a tensor in the range [0, 1] and sum to 1.
import torch input = torch.randn(5) print(input) softmax = torch.nn.Softmax(dim = 0) output = softmax(input) print(output) print(output.sum())
Output
tensor([-0.5654, -0.9031, -0.3060, -0.6847, -1.4268]) tensor([0.2315, 0.1651, 0.3001, 0.2055, 0.0978]) tensor(1.0000)
Notice that after rescaling, the elements of the tensor are in the range [0,1] and also the sum of elements of the rescaled tensor is 1.
Example 2
The following Python program rescales a tensor in the range [0, 1] and sum to 1.
# Import the required library import torch input = torch.randn(5,2) print(input) softmax = torch.nn.Softmax(dim = 1) output = softmax(input) print(output) print(output[0]) print(output[1].sum())
Output
tensor([[-0.5788, 0.9244], [-0.5172, 1.6231], [ 1.3032, -2.1107], [-0.4802, 0.1321], [-1.3219, -0.3570]]) tensor([[0.1819, 0.8181], [0.1052, 0.8948], [0.9681, 0.0319], [0.3515, 0.6485], [0.2759, 0.7241]]) tensor([0.1819, 0.8181]) tensor(1.)
Notice that after rescaling, the elements of the tensor are in the range [0,1] and also the sum of elements of rescaled tensor is 1.
2K+ Views
