Bidirectional LSTM in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the Bidirectional layer from Keras.

NLP

from tensorflow.keras.layers import [1]

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

ABidirectional

BDense

CConv2D

DDropout

Attempts:

3 left

💡 Hint

Common Mistakes

Importing Dense or Conv2D instead of Bidirectional.

Forgetting to import Bidirectional layer.

The Bidirectional layer wraps an RNN to make it process input in both forward and backward directions.

2fill in blank

medium

Complete the code to create a Bidirectional LSTM layer with 64 units.

NLP

model.add(Bidirectional([1](64)))

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

ADense

BLSTM

CGRU

DConv1D

Attempts:

3 left

💡 Hint

Common Mistakes

Using Dense or Conv1D inside Bidirectional, which is incorrect.

Using GRU instead of LSTM if the question specifically asks for LSTM.

The Bidirectional wrapper is used with recurrent layers like LSTM to process sequences forward and backward.

3fill in blank

hard

Fix the error in the code to correctly compile the model with categorical crossentropy loss.

NLP

model.compile(optimizer='adam', loss='[1]', metrics=['accuracy'])

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

Acategorical_crossentropy

Bmean_squared_error

Cbinary_crossentropy

Dhinge

Attempts:

3 left

💡 Hint

Common Mistakes

Using mean_squared_error for classification.

Using binary_crossentropy for multi-class problems.

For multi-class classification, categorical_crossentropy is the correct loss function.

4fill in blank

hard

Fill both blanks to create a Bidirectional LSTM layer that returns sequences and uses 128 units.

NLP

model.add(Bidirectional([1](128, [2]=True)))

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

ALSTM

Breturn_sequences

Cactivation

Dunits

Attempts:

3 left

💡 Hint

Common Mistakes

Using activation or units as keyword argument instead of return_sequences.

Not setting return_sequences when needed.

The LSTM layer inside Bidirectional should have return_sequences=True to output the full sequence.

5fill in blank

hard

Fill all three blanks to build a simple Bidirectional LSTM model for text classification.

NLP

model = Sequential()
model.add(Embedding(input_dim=[1], output_dim=[2], input_length=100))
model.add(Bidirectional(LSTM([3])))
model.add(Dense(5, activation='softmax'))

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

A10000

B64

C128

D32

Attempts:

3 left

💡 Hint

Common Mistakes

Confusing input_dim and output_dim values.

Using too small or too large numbers for embedding or LSTM units.

The Embedding layer needs input_dim=10000 for vocabulary size, output_dim=64 for embedding size, and the LSTM layer uses 128 units inside Bidirectional.