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 tf.layers.embedding() function is used to map positive integers into dense vectors of fixed size.
Syntax:
tf.layers.embedding(args)
Parameters: This function accepts the args as a parameter which can have the following properties:
- inputDim: It is used to specify the vocabulary size.
- outputDim: It is used to specify the dimension of the dense embedding.
- embeddingsInitializer: It is used to specify the initializer of the embeddings matrix.
- embeddingsRegularizer: It is used to specify which regularizer function is applied to embeddings matrix.
- activityRegularizer: It is used to specify which regularizer function is applied to activation.
- embeddingsConstraint: It is used to specify which constraint function is applied to the embeddings matrix.
- maskZero: It is used to check whether the input value 0 is a special padding value.
- inputLength: It is used to specify the length of input sequences.
- inputShape: It is used to create an input layer to insert before this layer.
- batchInputShape: It is used to create an input layer to insert before this layer.
- batchSize: It is used to construct the batchInputShape if the inputShape is specified and batchInputShape is not specified.
- dtype: It is used to denote the data-type for this layer.
- name: It is used to denote the name for this layer.
- trainable: It is used to indicate whether the weights of this layer are updatable by fit or not.
- weights: It is used to denote the initial weight values of the layer.
- inputDType: It is just for the legacy support and not to be use for new code.
Return value: It returns the Embedding.
Example 1:
Javascript
// Import library import * as tf from "@tensorflow/tfjs" // Create embedding layer const embeddingLayer = tf.layers.embedding({ inputDim: 10, outputDim: 3, inputLength: 2 }); const input = tf.ones([2, 2]); // Apply embedding to input const output = embeddingLayer.apply(input); // Print the output console.log(output) |
Output:
Tensor
[[[0.0179072, 0.0069226, 0.0202718],
[0.0179072, 0.0069226, 0.0202718]],
[[0.0179072, 0.0069226, 0.0202718],
[0.0179072, 0.0069226, 0.0202718]]]
Example 2:
Javascript
// Import the library import * as tf from "@tensorflow/tfjs" // Create embedding layer const embeddingLayer = tf.layers.embedding({ inputDim: 100, outputDim: 4, inputLength: 3 }); const input = tf.ones([3, 3]); // Apply embedding to input const output = embeddingLayer.apply(input); // Print the output console.log(output) |
Output:
Tensor
[[[0.0443502, -0.0342815, 0.0228792, 0.0198386],
[0.0443502, -0.0342815, 0.0228792, 0.0198386],
[0.0443502, -0.0342815, 0.0228792, 0.0198386]],
[[0.0443502, -0.0342815, 0.0228792, 0.0198386],
[0.0443502, -0.0342815, 0.0228792, 0.0198386],
[0.0443502, -0.0342815, 0.0228792, 0.0198386]],
[[0.0443502, -0.0342815, 0.0228792, 0.0198386],
[0.0443502, -0.0342815, 0.0228792, 0.0198386],
[0.0443502, -0.0342815, 0.0228792, 0.0198386]]]
Reference: https://js.tensorflow.org/api/latest/#layers.embedding
Whether you’re 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 we’ve already empowered, and we’re here to do the same for you. Don’t miss out – check it out now!
