Tensorflow.js tf.LayersModel class .getLayer() Method

Last Updated : 04 Aug, 2021

Tensorflow.js is an open-source library that is developed by Google for running machine learning models as well as deep learning neural networks in the browser or node environment.

The .getLayer() function is used to fetch a layer that is based upon either its name (which must be unique) or else an index. Where, indices are relying on the order of the horizontal graph traversal in bottom-up fashion. Moreover, in case both name as well as index are given, then the index will take priority.

Syntax:

getLayer(name?, index?)

Parameters:

name: It is the stated name of the layer. It is optional and is of type string.
index: It is the stated index of the layer. It is optional and is of type number.

Return Value: It returns tf.layers.Layer.

Example 1:

JavaScript

// Importing the tensorflow.js library
import * as tf from "@tensorflow/tfjs"

// Defining model
const model = tf.sequential();

// Adding a layer
model.add(tf.layers.dense({units: 4, inputShape: [1]}));

// Calling getLayer() method
const layer_0 = model.getLayer(null, 0);

// Printing weights of the layer_0
// using getWeights() method
layer_0.getWeights()[0].print();

Output:

Tensor
     [[-0.0678914, 0.6647689, -0.3708572, -0.1764591],]

Example 2:

JavaScript

// Importing the tensorflow.js library
import * as tf from "@tensorflow/tfjs"

// Defining model
const model = tf.sequential();

// Adding layers
model.add(tf.layers.dense({units: 4, inputShape: [1]}));
model.add(tf.layers.dense({units: 2, inputShape: [3], activation: 'relu6'}));
model.add(tf.layers.dense({units: 3, inputShape: [5], activation: 'sigmoid'}));

// Calling getLayer() method
const layer_0 = model.getLayer(NaN, 0);
const layer_1 = model.getLayer('denselayer', 1);
const layer_2 = model.getLayer(undefined, 2);

// Printing number of numbers in the weights
// of the layer_0, layer_1, and layer_2 
// using countParams() method
console.log(layer_0.countParams());
console.log(layer_1.countParams());
console.log(layer_2.countParams());