Tensorflow.js tf.layers.bidirectional() Function

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

The tf.layers.bidirectional function is a bidirectional wrapper for RNNs layer.

Syntax:

tf.layers.bidirectional( args )

Parameters: This function accepts objects as parameters with the following fields:

• layers: It is the instance of RNN layer which is to be wrapped.
• mergeMode: It should be ‘sum’ or ‘mul’ or ‘concat’ or ‘ave’. It is the mode at which the output of forward and backward or RNNs are combined.
• inputShape:It should be an array of numbers. This field is used to create the input layer which is used to be inserted before this layer.
• batchInputShape: It should be an array of numbers. This field will be used if the input shape and this field are provided as a parameter for creating the input layer which is used to insert before this layer.
• batchSize: It should be a number. In the absence of batchInputShape, this field is used to create batchInputShape with input shape. batchInputShape : [ batchSize ,  …inputShape].
• dtype: If this layer is used as the input layer, then this field is used as the data type for this layer.
• name: It should be a string type. this field defines the name for this layer.
• trainable: It should be boolean. This field defines whether the weights of this layer are trainable with fit or not.
• weights: This should be a tensor that defines the initial weight value for this layer.

Returns: It returns Bidirectional.

Example 1:

Javascript

import * as tf from '@tensorflow/tfjs'

// Bidirectional layer
const k = tf.layers.bidirectional({
   layer: tf.layers.dense({units: 4}),
   batchInputShape: [32, 10, 16],
});

// Creating input layer
const input = tf.input({shape: [10,16,32]});
const output = k.apply(input);

// Printing the Input Shape
console.log(JSON.stringify(output.shape));

Output:

[null,10,16,8]

Example 2:

Javascript

import * as tf from '@tensorflow/tfjs'

// Instance of RNN layer
const RNN = tf.layers.dense({units: 3});
// Creating Bidirectional layer
const k = tf.layers.bidirectional({
   layer: RNN,
   mergeMode: 'sum',
   batchInputShape: [8, 4, 2],
});

// Create a 3d tensor
const x = tf.tensor([1, 2, 3, 4], [4, 1]);

// Apply Bidirectional layer to x
const output = k.apply(x);

// Print output
output.print()

Output:

Tensor
    [[-0.6737164, -1.6011676, 1.9193256],
     [-1.3474327, -3.2023351, 3.8386512],
     [-2.0211492, -4.8035026, 5.7579765],
     [-2.6948655, -6.4046702, 7.6773024]]

Reference: https://js.tensorflow.org/api/latest/#layers.bidirectional