Tensorflow.js tf.fill() Function

Tensorflow.js is an open-source library for creating machine learning models in Javascript that allows users to run the models directly in the browser.

The tf.fill() is a function defined in the class tf.Tensor. It is used to create a tensor that is filled with a scalar value.

Syntax:

tf.fill( shape, value, dtype )

Parameters:

• shape: It's an array of integers defining the shape of the output tensor.
• value: It's a scalar value with which the output tensor is to be filled.
• dtype: It defines the data type of elements in the output tensor. It can be 'float32'|'int32'|'bool'|'complex64'|'string'. It's optional to include and default value is 'float32'.

Return value: It returns the tensor of a specified shape filled with a scalar value.

Example 1: Filling the tensor with a scalar number

• Creating a tensor of shape [4, 2] filled with scalar value 2.
• It takes the default datatype of elements in the output tensor as float.

// Dynamic loading the "@tensorflow/tfjs" module
const tf = require('@tensorflow/tfjs');
require('@tensorflow/tfjs-node');

// Creating a tensor of of shape [4,2] filled with 
// scalar value 2
var matrix = tf.fill(shape = [4,2],value = 2)
  
// Printing the tensor
matrix.print()

Output:

Tensor
    [[2, 2],
     [2, 2],
     [2, 2],
     [2, 2]]

Example 2: Explicitly defining the data type of elements

• Create a tensor of shape [3, 4] filled with string 'Gfg'. 

// Dynamic loading the "@tensorflow/tfjs" module
const tf = require('@tensorflow/tfjs');
require('@tensorflow/tfjs-node');

// Creating a tensor  of shape [3,4] filled
// with string value 'Gfg'
var matrix = tf.fill(shape = [3, 4], 
        value = 'Gfg', dtype = 'string')
  
// Printing the tensor
matrix.print()

Output:

Tensor
    [['Gfg', 'Gfg', 'Gfg', 'Gfg'],
     ['Gfg', 'Gfg', 'Gfg', 'Gfg'],
     ['Gfg', 'Gfg', 'Gfg', 'Gfg']]