Bidirectional RNN in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a bidirectional RNN layer using TensorFlow.

TensorFlow

import tensorflow as tf

model = tf.keras.Sequential([
    tf.keras.layers.[1](
        tf.keras.layers.SimpleRNN(32), input_shape=(10, 8))
])

Drag options to blanks, or click blank then click option'

ABidirectional

BDense

CConv1D

DDropout

Attempts:

3 left

💡 Hint

Common Mistakes

Using Dense or Conv1D layers instead of Bidirectional for RNNs.

Not wrapping the RNN layer inside the Bidirectional wrapper.

The Bidirectional wrapper is used to create a bidirectional RNN layer in TensorFlow.

2fill in blank

medium

Complete the code to get the output shape of a bidirectional RNN layer.

TensorFlow

import tensorflow as tf

layer = tf.keras.layers.Bidirectional(tf.keras.layers.SimpleRNN(16))
input_data = tf.random.normal([5, 10, 8])
output = layer(input_data)
print(output.shape)  # Expected shape: (5, [1])

Drag options to blanks, or click blank then click option'

A16

B32

C8

D10

Attempts:

3 left

💡 Hint

Common Mistakes

Expecting output shape to be (5, 16) instead of (5, 32).

Confusing batch size with feature size.

The output size doubles because the RNN runs forward and backward, so 16 units become 32.

3fill in blank

hard

Fix the error in the code to correctly create a bidirectional LSTM layer.

TensorFlow

import tensorflow as tf

model = tf.keras.Sequential([
    tf.keras.layers.Bidirectional(tf.keras.layers.[1](64), input_shape=(20, 10))
])

Drag options to blanks, or click blank then click option'

ALSTM

BDense

CSimpleRNN

DConv2D

Attempts:

3 left

💡 Hint

Common Mistakes

Using Dense or Conv2D layers inside Bidirectional.

Using SimpleRNN when LSTM is required.

The LSTM layer is the correct RNN type to use inside the Bidirectional wrapper for LSTM networks.

4fill in blank

hard

Fill both blanks to create a bidirectional GRU layer that returns sequences.

TensorFlow

import tensorflow as tf

layer = tf.keras.layers.Bidirectional(
    tf.keras.layers.[1](128, return_sequences=[2])
)

Drag options to blanks, or click blank then click option'

AGRU

BTrue

CFalse

DSimpleRNN

Attempts:

3 left

💡 Hint

Common Mistakes

Setting return_sequences to False when sequences are needed.

Using SimpleRNN instead of GRU.

Use GRU for the layer type and set return_sequences=True to get output at each time step.

5fill in blank

hard

Fill all three blanks to build a bidirectional RNN model with an LSTM layer, a dropout layer, and a dense output layer.

TensorFlow

import tensorflow as tf

model = tf.keras.Sequential([
    tf.keras.layers.Bidirectional(tf.keras.layers.[1](64), input_shape=(15, 12)),
    tf.keras.layers.[2](0.5),
    tf.keras.layers.[3](10, activation='softmax')
])

Drag options to blanks, or click blank then click option'

ALSTM

BDropout

CDense

DSimpleRNN

Attempts:

3 left

💡 Hint

Common Mistakes

Using SimpleRNN instead of LSTM for the first layer.

Forgetting to add Dropout for regularization.

Using Dense without activation or wrong activation.

The model uses a LSTM inside Bidirectional, followed by a Dropout layer to reduce overfitting, and a Dense layer with softmax activation for classification output.