Tensorflow.js tf.train.Optimizer class .applyGradients() Method
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.
Tensorflow.js tf.train.Optimizer .apply Gradients( ) is used for Updating variables by using the computed gradients.
Syntax:
Optimizer.applyGradients( variableGradients );
Parameters:
- variableGradients( { [ name : String ] : tf.Tensor } | NamedTensor[ ]): A mapping of variable name to its gradients value.
Returns: void
Example 1: In this example, we will updates the value of variable with the help of applyGradients( ) method of the default value optimizer.
Javascript
import * as tf from "@tensorflow/tfjs"
const xs = tf.tensor1d([0, 1, 2]);
const ys = tf.tensor1d([1.58, 2.24, 3.41]);
const x = tf.scalar(Math.random()).variable();
const y = tf.scalar(Math.random()).variable();
const f = x => (x.square()).add(y);
const learningRate = 0.05;
const optimizer =
tf.train.rmsprop(learningRate);
const varGradients = f(xs).dataSync();
for (let i = 0; i < 5; i++){
optimizer.applyGradients(varGradients);
}
console.log(
`x: ${x.dataSync()}, y: ${y.dataSync()}`);
const preds = f(xs).dataSync();
preds.forEach((pred, i) => {
console.log(`x: ${i}, pred: ${pred}`);
});
|
Output:
x: -0.526353657245636, y: 0.15607579052448273
x: 0, pred: 0.15607579052448273
x: 1, pred: 1.1560758352279663
x: 2, pred: 4.156075954437256
Example 2: In this example, we will update the variable with the help of applyGradients( ) method of custom optimizer.
Javascript
import * as tf from "@tensorflow/tfjs"
const xs = tf.tensor1d([0, 1, 2, 3]);
const ys = tf.tensor1d([1.3, 3.7, 12.4, 26.6]);
const a = tf.scalar(Math.random()).variable();
const b = tf.scalar(Math.random()).variable();
const c = tf.scalar(Math.random()).variable();
const f = x => a.mul(x.mul(3)).add(b.square(x)).add(c);
const learningRate = 0.01;
const initialAccumulatorValue = 10;
const optimizer = tf.train.adagrad(learningRate,
initialAccumulatorValue);
const varGradients = f(xs).dataSync();
for (let i = 0; i < 8; i++){
optimizer.applyGradients(varGradients)}
console.log(`a: ${a.dataSync()},
b: ${b.dataSync()}, c: ${c.dataSync()}`);
const preds = f(xs).dataSync();
preds.forEach((pred, i) => {
console.log(`x: ${i}, pred: ${pred}`);
});
|
Output:
a: 0.032658617943525314,
b: 0.9213025569915771, c: 0.7167015671730042
x: 0, pred: 1.565500020980835
x: 1, pred: 1.663475751876831
x: 2, pred: 1.7614517211914062
x: 3, pred: 1.8594274520874023
Reference: https://js.tensorflow.org/api/3.8.0/#tf.train.Optimizer.applyGradients
Last Updated :
03 Sep, 2021
Like Article
Save Article
Share your thoughts in the comments
Please Login to comment...