Skip to content
Related Articles

Related Articles

Save Article
Improve Article
Save Article
Like Article

How to Get the Data Type of a Pytorch Tensor?

  • Last Updated : 21 Jul, 2021

In this article, we are going to create a tensor and get the data type. The Pytorch is used to process the tensors. Tensors are multidimensional arrays. PyTorch accelerates the scientific computation of tensors as it has various inbuilt functions.

Vector:

A vector is a one-dimensional tensor that holds elements of multiple data types. We can create a vector using PyTorch. Pytorch is available in the Python torch module so, we need to import it

 Attention geek! Strengthen your foundations with the Python Programming Foundation Course and learn the basics.  

To begin with, your interview preparations Enhance your Data Structures concepts with the Python DS Course. And to begin with your Machine Learning Journey, join the Machine Learning - Basic Level Course

Syntax:



import pytorch

Creation of One-Dimensional Tensors:

One dimensional vector is created using the torch.tensor() method.

Syntax:

torch.tensor([element1,element2,.,element n],dtype)

Parameters:

  • dtype: Specify the data type.
dtype=torch.datatype

Example: Python program to create tensor elements not specifying the data type.

Python3




# importing torch module
import torch
  
# create one dimensional tensor with
# integer type elements
a = torch.tensor([10, 20, 30, 40, 50])
print(a)
  
# create one dimensional tensor with 
# float type elements
b = torch.tensor([10.12, 20.56, 30.00, 40.3, 50.4])
print(b)

Output:

tensor([10, 20, 30, 40, 50])
tensor([10.1200, 20.5600, 30.0000, 40.3000, 50.4000])

Supported Data Types:

The following data types are supported by vector:

Data TypeDescription
int8Integer type with 8 bytes
uint8Unsigned integer type with 8 bytes
int16Integer type with 16 bytes
int32Integer type with 32 bytes
int64Integer type with 64 bytes
floatData with float type(decimal)
doubleData with float type (64 bit) decimal
boolBoolean type: returns True if the value is greater than 0, otherwise False

We can get the data type by using dtype command:

Syntax:



tensor_name.dtype

Example 1: Python program to create tensor with integer data types and display data type

Python3




# import torch
import torch
  
  
# create a tensor with unsigned integer type of 8 bytes size
a = torch.tensor([100, 200, 2, 3, 4], dtype=torch.uint8)
# display tensor
print(a)
# display data type
print(a.dtype)
  
# create a tensor with  integer type of 8 bytes size
a = torch.tensor([1, 2, -6, -8, 0], dtype=torch.int8)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)
  
# create a tensor with  integer type of 16 bytes size
a = torch.tensor([1, 2, -6, -8, 0], dtype=torch.int16)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)
  
  
# create a tensor with  integer type of 32 bytes size
a = torch.tensor([1, 2, -6, -8, 0], dtype=torch.int32)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)
  
# create a tensor with  integer type of 64 bytes size
a = torch.tensor([1, 2, -6, -8, 0], dtype=torch.int64)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)

Output:

tensor([100, 200,   2,   3,   4], dtype=torch.uint8)
torch.uint8
tensor([ 1,  2, -6, -8,  0], dtype=torch.int8)
torch.int8
tensor([ 1,  2, -6, -8,  0], dtype=torch.int16)
torch.int16
tensor([ 1,  2, -6, -8,  0], dtype=torch.int32)
torch.int32
tensor([ 1,  2, -6, -8,  0])
torch.int64

Example 2: Create float type and display data types.

Python3




# import torch
import torch
  
  
# create a tensor with  float type
a = torch.tensor([100, 200, 2, 3, 4], dtype=torch.float)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)
  
# create a tensor with  double type
a = torch.tensor([1, 2, -6, -8, 0], dtype=torch.double)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)

Output:

tensor([100., 200.,   2.,   3.,   4.])
torch.float32
tensor([ 1.,  2., -6., -8.,  0.], dtype=torch.float64)
torch.float64

Example 3: Create a tensor with boolean type

Python3




# import torch
import torch
  
  
# create a tensor with  bool type
a = torch.tensor([100, 200, 2, 3, 4], dtype=torch.bool)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)
  
# create a tensor with  bool type
a = torch.tensor([0, 0, 0, 1, 2], dtype=torch.bool)
  
# display tensor
print(a)
  
# display data type
print(a.dtype)

Output:

tensor([True, True, True, True, True])
torch.bool
tensor([False, False, False,  True,  True])
torch.bool



My Personal Notes arrow_drop_up
Recommended Articles
Page :