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.customGrad() function is used to return the gradient of a specified custom function âfâ. Here the custom function gives {value: Tensor, gradFunc: (dy, saved) â Tensor[]}, where gradFunc gives the custom gradients of the input function f in respect of its inputs.
Syntax:
tf.customGrad(f)
Parameters: This function accepts a parameter which is illustrated below:
- f: It is the specified custom function.
Return Value: This function returns the gradient of a specified custom function âfâ
Example 1:
Javascript
// Importing the tensorflow.js libraryimport * as tf from "@tensorflow/tfjs"Â
// Initializing a custom function fconst f = (a, save) => {      // Saving a for its availability later for the gradient   save([a]);      // Overriding gradient of a^2   return {     value: a.square(),           // Here "saved.a" pointing to "a" which     // have been saved above     gradFunc: (dy, saved) => [dy.mul(saved[0].abs())]   };}Â
// Calling the .customGrad() function // over the above specified custom function fconst customOp = tf.customGrad(f);Â
// Initializing a 1D Tensor of some valuesconst a = tf.tensor1d([0, -1, 1, 2]);Â
// Getting the gradient of above function// f for the above specified Tensor valuesconst da = tf.grad(a => customOp(a));Â
// Printing the custom function "f" for the// above specified Tensor "a"console.log(`f(a):`);customOp(a).print();Â
// Printing the gradient of the function "f" for the// above specified Tensor "a"console.log(`f'(a):`);da(a).print(); |
Output:
f(a): Tensor [0, 1, 1, 4] f'(a): Tensor [0, 1, 1, 2]
Example 2:
Javascript
// Importing the tensorflow.js libraryimport * as tf from "@tensorflow/tfjs"Â
// Calling the .customGrad() function // with the custom function "f" as // it's parameterconst customOp = tf.customGrad(   // Initializing a custom function f(a, save) => {      // Saving a for its availability later for the gradient   save([a]);      // Overriding gradient of a^3   return {     value: a.pow(tf.scalar(3, 'int32')),           // Here "saved.a" pointing to "a" which     // have been saved above     gradFunc: (dy, saved) => [dy.mul(saved[0].abs())]   };});Â
// Initializing a 1D Tensor of some valuesconst a = tf.tensor1d([0, -1, 2, -2, 0.3]);Â
// Getting the gradient of above function// f for the above specified Tensor valuesconst da = tf.grad(a => customOp(a));Â
// Printing the custom function "f" for the// above specified Tensor "a"console.log(`f(a):`);customOp(a).print();Â
// Printing the gradient of the function "f" for the// above specified Tensor "a"console.log(`f'(a):`);da(a).print(); |
Output:
f(a): Tensor [0, -1, 8, -8, 0.027] f'(a): Tensor [0, 1, 2, 2, 0.3]
Reference:https://js.tensorflow.org/api/latest/#customGrad
Â
Whether youâre preparing for your first job interview or aiming to upskill in this ever-evolving tech landscape, neveropen Courses are your key to success. We provide top-quality content at affordable prices, all geared towards accelerating your growth in a time-bound manner. Join the millions weâve already empowered, and weâre here to do the same for you. Donât miss out â check it out now!
