Open In App

Tensorflow.js tf.losses.huberLoss() Function

Improve
Improve
Like Article
Like
Save
Share
Report

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. It also helps the developers to develop ML models in JavaScript language and can use ML directly in the browser or in Node.js.

The Tensorflow.js tf.losses.huberLoss() function calculates the Huber loss between two given tensors.

Syntax:

tf.losses.huberLoss(
    labels, predictions,
    weights, delta, reduction
);

Parameters:

  • labels: It is the ground truth output tensor. It is similar in dimensions to ‘predictions‘.
  • predictions: It is the outputs that are being predicted.
  • weights: These are those tensors whose rank is either 0 or 1, and they must be broadcastable to loss of shape.
  • delta: It is that point throughout where huberLoss converts to linear from quadratic.
  • reduction: It is the type of reduction to apply to loss. It must be of Reduction type.

Note: The weights, delta, and reduction are optional parameters.

Return value: It returns tf.Tensor.

Example 1:

Javascript




// Importing the tensorflow.Js library
import * as tf from "@tensorflow/tfjs"
 
// Initializing tensor1 as geek1
let geek1 = tf.tensor2d([[1, 2, 5], [6, 7, 10]]);
 
// Initializing tensor2 as geek2
let geek2 = tf.tensor2d([[5, 7, 11], [2, 4, 8]])
 
// Computing huber loss between geek1 and geek2
// using .huberLoss() function
geek = tf.losses.huberLoss(geek1, geek2)
geek.print();


Output: 

Tensor
    3.5

Example 2: 

Javascript




// Importing the tensorflow.Js library
import * as tf from "@tensorflow/tfjs"
 
// Computing huber loss between two 3D
// tensors and printing the result
tf.losses.huberLoss(
    tf.tensor4d([[[[9], [8]], [[7], [5]]]]),
    tf.tensor4d([[[[1], [2]], [[3], [4]]]])
).print();


Output: 

Tensor
    4.25

Reference: https://js.tensorflow.org/api/latest/#losses.huberLoss

 



Last Updated : 09 Sep, 2021
Like Article
Save Article
Previous
Next
Share your thoughts in the comments
Similar Reads