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 tf.whereAsync() function is used to return the 2-D tensor of coordinates of true elements for the specified condition. In the returned tensor, 1st dimension i.e. rows specifies the number of true elements, and the 2nd dimension i.e. columns specifies the co-ordinates of the true elements i.e. the output tensor is having the shape of [numTrueElems, condition.rank].
Note: The shape of the returned tensor depends on the true values present in the input.
Syntax:
tf.whereAsync (condition)
Parameters: This function accepts a parameter which is illustrated below:
- condition: The input condition is array of Boolean values.
Return Value: It returns the 2-D tensor of coordinates of true elements for the specified condition.
Example 1:
Javascript
// Importing the tensorflow.js library import * as tf from "@tensorflow/tfjs" // Initializing a tensor for conditions const cond = tf.tensor1d([false, true, true, false, true], 'bool'); // Calling the .whereAsync() function const result = await tf.whereAsync(cond); // Getting the 2-D tensor of true elements result.print(); |
Output:
Tensor
[[1],
[2],
[4]]
Example 2:
Javascript
// Importing the tensorflow.js library import * as tf from "@tensorflow/tfjs" // Using a tensor of conditions as the // parameter of the .whereAsync() function const result = await tf.whereAsync(tf.tensor1d( [true, false, true, true, false], 'bool')); // Getting the 2-D tensor of true elements result.print(); |
Output:
Tensor
[[0],
[2],
[3]]
Reference:https://js.tensorflow.org/api/1.0.0/#whereAsync
