Tensorflow.js is an open-source library that is being developed by Google for running machine learning models as well as deep learning neural networks in the browser or node environment.
The .unique() function is used to find unique elements along an axis of a tensor.
Syntax:
tf.unique (x, axis?)
Parameters :
- x: It is a first tensor input that can be of type tf.Tensor, or TypedArray, or Array. Parameters can be of these types (int32, string, bool).
- axis ( number ): It is a second tensor input which is optional. The axis of the tensor is used to find the unique elements.
Return value : It returns the values and indices.
Example 1: In this example, we are defining an input tensor of integer type and then printing the unique value of it. For creating an input tensor, we are utilizing the .tensor1d() method and in order to print the output we are using the .print() method.
Javascript
// Importing the tensorflow.js library import * as tf from "@tensorflow/tfjs" // Defining tensor input elements const a = tf.tensor1d([1, 1, 2, 2, 4, 4, 7, 7, 8]); // Calling unique() method const {values, indices} = tf.unique(a); // Printing the values and indices values.print(); // [1, 2, 4, 7, 8,] indices.print(); // [0, 0, 1, 1, 2, 2, 3, 3, 4] |
Output:
Tensor
[1, 2, 4, 7, 8]
Tensor
[0, 0, 1, 1, 2, 2, 3, 3, 4]
Example 2: In this example axis parameter has been used. For creating an input tensor, we are utilizing the .tensor2d() method and in order to print the output we are using the .print() method.
Javascript
// Importing the tensorflow.js library import * as tf from "@tensorflow/tfjs" // Defining tensor input elements const a = tf.tensor2d([[0, 0, 0], [2, 0, 0], [2, 0, 0]]); // Calling the unique() method const {values, indices} = tf.unique(a, 0) // Printing the values and indices values.print(); // [[1, 0, 0], // [2, 0, 0]] indices.print(); // [0, 0, 1] |
Output:
Tensor
[[0, 0, 0],
[2, 0, 0]]
Tensor
[0, 1, 1]
Reference: https://js.tensorflow.org/api/latest/#unique
