| by Arround The Web | No comments

How to if the Object is a PyTorch Tensor and Return the Metadata of a Tensor in PyTorch?

“In this PyTorch tutorial, we will see how to get the information from the given tensor in PyTorch.

PyTorch is an open-source framework available with a Python programming language.

A tensor is a multidimensional array that is used to store the data. So for using a Tensor, we have to import the torch module.

It is possible to check whether the given object is a tensor or not.

torch.is_tensor() is used to check whether the given object is tensor or not.

If the object is a tensor, it will return True otherwise, False.”

Syntax:

torch.is_tensor(object)

Parameter:

object refers to the collection of data.

Example 1

Here, we will create a tensor with 5 elements and check if it is a tensor or not.

#import torch module

import torch

 

#create a 1D tensor with 5 elements

data1 = torch.tensor([23,45,67,0,0])

 

#check whether data1 is tensor or not

print(torch.is_tensor(data1))

Output:

True

We can see that the given object is a tensor. So it returned True.

Example 2

Let’s create a list with 5 elements and check if it is tensor or not.

#import torch module

import torch

 

#create a list with 5 elements

data1 = [23,45,67,0,0]

 

 

#check whether data1 is tensor or not

print(torch.is_tensor(data1))

Output:

False

It returned False.

Now, we will see how to return the metadata of a tensor.

Metadata explains the tensor structure and elements present in the vector.

torch.size()

torch.size() returns the total number of elements present in a tensor.

Syntax:

tensor_object.size()

Where tensor_object is the tensor.

It takes no parameters.

Example 1

Let’s create a 1D tensor and return size.

#import torch module

import torch

 

#create a 1D tensor with 5 elements

data1 = torch.tensor([23,45,67,0,0])

 

#display

print("Tensor: ",data1)

#return tensor size

print("Size: ",data1.size())

Output:

Tensor: tensor([23, 45, 67, 0, 0])

Size: torch.Size([5])

We can see that 5 is returned since there are 5 elements in the above tensor.

Example 2

Let’s create a 2D tensor and return size.

#import torch module

import torch

 

#create a 2D tensor with 5 elements in each row

data1 = torch.tensor([[23,45,67,0,0],[23,45,67,0,0]])

 

#display

print("Tensor: ",data1)

#return tensor size

print("Size: ",data1.size())

Output:

Tensor: tensor([[23, 45, 67, 0, 0],

[23, 45, 67, 0, 0]])

Size: torch.Size([2, 5])

We can see that 2,5 is returned and represents 2 rows and 5 columns.

torch.shape

torch.shape() returns the shape of a tensor.

Syntax:

tensor_object.shape

Where tensor_object is the tensor.

It takes no parameters.

Example 1

#import torch module

import torch

 

#create a 1D tensor with 5 elements

data1 = torch.tensor([23,45,67,0,0])

 

#display

print("Tensor: ",data1)

 

#return tensor Shape

print("Shape: ",data1.shape)

Output:

Tensor: tensor([23, 45, 67, 0, 0])

Shape: torch.Size([5])

We can see that 5 is returned since there are 5 elements in the above tensor.

Example 2

#import torch module

import torch

 

#create a 2D tensor with 5 elements in each row

data1 = torch.tensor([[23,45,67,0,0],[23,45,67,0,0]])

 

#display

print("Tensor: ",data1)

 

#return tensor Shape

print("Shape: ",data1.shape)

Output:

Tensor: tensor([[23, 45, 67, 0, 0],

[23, 45, 67, 0, 0]])

Shape: torch.Size([2, 5])

We can see that 2,5 is returned and represents 2 rows and 5 columns.

torch.numel()

torch.numel() returns the total number of elements present in a tensor.

Syntax:

tensor_object.numel()

Where tensor_object is the tensor.

It takes no parameters.

Example 1

#import torch module

import torch

 

#create a 1D tensor with 5 elements

data1 = torch.tensor([23,45,67,0,0])

 

#display

print("Tensor: ",data1)

 

#return total number of elements in a tensor

print("Total elements: ",data1.numel())

Output:

Tensor: tensor([23, 45, 67, 0, 0])

Total elements: 5

We can see that 5 is returned since there are 5 elements in the above tensor.

Example 2

#import torch module

import torch

 

 

#create a 2D tensor with 5 elements in each row

data1 = torch.tensor([[23,45,67,0,0],[23,45,67,0,0]])

 

#display

print("Tensor: ",data1)

 

#return total number of elements in a tensor

print("Total elements: ",data1.numel())

Output:

Tensor: tensor([[23, 45, 67, 0, 0],

[23, 45, 67, 0, 0]])

Total elements: 10

We can see that 10 is returned since there are a total of 10 elements present in the tensor.

Conclusion

In this PyTorch lesson, we saw how to check if the given object is tensor or not using the is_tensor() function. To return the metadata, we used size() and shape methods to return the size and shape of the given tensor.

Share Button

Source: linuxhint.com

Leave a Reply