Tensorflow.js tf.scalar() Function

Tensorflow.js is an open-source library developed by Google for running machine learning models and deep learning neural networks in the browser or node environment.

The .scalar() function is used to create a scalar type of tensor means. A scalar is a zero-dimension array and is also called a rank-0 Tensor. A scalar is created using .scalar() function.

Syntax:

t.scalar( value, dataType )

Parameters:

• value: The value of the scalar. The value can be a number, string, Uint8Array[ ], or boolean.
• dataType [Optional]: The data type of the value. It can be an int32, float32, bool, complex64, or string.

Return Value: It returns the Tensor Object.

Creating a Scalar: In this example, we are creating a new scalar, which means a tensor of only one value.

Example 1:

// Importing the tensorflow.js library
import * as tf from "@tensorflow/tfjs"
// Value of a scalar
var value = 12
// Creating the value of a scalar
var tens = tf.scalar(value)
// Printing the scalar
tens.print();

 
 Output:

Tensor
    12

Creating a scalar of a specific data type: In this example, we are creating a scalar of a specific data type. Note that the data type should only int32, float32, bool, complex64, or string.

Example 2:

// Importing the tensorflow.js library
import * as tf from "@tensorflow/tfjs"
// Creating a scalar using int value
var int_tensor = tf.scalar(12, 'int32')
int_tensor.print()
// Creating a scalar using string value
var str_tensor = tf.scalar("GFG", "string")
str_tensor.print()
// Creating a scalar using float value
var float_tensor = tf.scalar(12.6, "float32")
float_tensor.print();
// Creating a scalar using bool value
var bool_tensor1 = tf.scalar(true, "bool")
bool_tensor1.print()
// Creating a scalar using bool(0 and 1) type 
var bool_tensor2 = tf.scalar(0, "bool")
bool_tensor2.print()

Output:

Tensor
    12
Tensor
    GFG
Tensor
    12.600000381469727
Tensor
    true
Tensor
    false

Note: You can also create a scalar using tf.tensor() function. Let's see the example 

Creating a Scalar using tf.tensor() Function: 

Example 3:

// Importing the tensorflow.js library
import * as tf from "@tensorflow/tfjs"
// Creating a scalar using int tf.tensor()
var tens = tf.tensor(12, [], "int32")
tens.print()

Here we are providing the second parameter of the function an empty array because we are creating a scalar and a scalar is a rank-0 tensor. 

Output:

Tensor
    12