Embedding layer usage in NLP - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create an embedding layer with input dimension 1000 and output dimension 64.

NLP

embedding = nn.Embedding([1], 64)

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

A10

B64

C128

D1000

Attempts:

3 left

💡 Hint

Common Mistakes

Using the output dimension as the input dimension

Confusing input and output dimensions

The input dimension is the size of the vocabulary, which is 1000 here.

2fill in blank

medium

Complete the code to convert input word indices to embeddings using the embedding layer.

NLP

embedded_words = embedding([1])

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

Aword_indices

Bembedding

Cinput_dim

Doutput_dim

Attempts:

3 left

💡 Hint

Common Mistakes

Passing the embedding layer itself as input

Passing dimensions instead of indices

You pass the word indices tensor to the embedding layer to get embeddings.

3fill in blank

hard

Fix the error in the embedding layer initialization by filling the correct padding index.

NLP

embedding = nn.Embedding(1000, 64, padding_idx=[1])

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

A-1

B0

C64

D1000

Attempts:

3 left

💡 Hint

Common Mistakes

Using -1 which is invalid

Using output dimension as padding index

The padding index is usually 0, which tells the embedding layer to ignore that index during training.

4fill in blank

hard

Fill both blanks to create an embedding layer and get embeddings for input indices.

NLP

embedding = nn.Embedding([1], [2])
embedded = embedding(input_indices)

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

A5000

B128

C64

D1000

Attempts:

3 left

💡 Hint

Common Mistakes

Swapping input and output dimensions

Using wrong vocabulary size

The embedding layer is created with input dimension 1000 and output dimension 64.

5fill in blank

hard

Fill all three blanks to create an embedding layer, set padding index, and get embeddings.

NLP

embedding = nn.Embedding([1], [2], padding_idx=[3])
embedded = embedding(batch_indices)

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

A10000

B300

C0

D1

Attempts:

3 left

💡 Hint

Common Mistakes

Using invalid padding index

Confusing embedding size with vocabulary size

The embedding layer uses vocabulary size 10000, embedding size 300, and padding index 0.