nn.LSTM layer in PyTorch - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create an LSTM layer with input size 10 and hidden size 20.

PyTorch

lstm = nn.LSTM(input_size=[1], hidden_size=20)

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

A5

B10

C20

D15

Attempts:

3 left

💡 Hint

Common Mistakes

Confusing input_size with hidden_size.

Using a value that does not match the input data features.

The input_size parameter defines the number of expected features in the input. Here, it should be 10.

2fill in blank

medium

Complete the code to initialize hidden and cell states for an LSTM with batch size 3 and hidden size 5.

PyTorch

h0 = torch.zeros(1, [1], 5)
c0 = torch.zeros(1, [1], 5)

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

A10

B1

C5

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Using hidden size instead of batch size for the second dimension.

Confusing the order of dimensions.

The second dimension is the batch size, which should be 3 here.

3fill in blank

hard

Fix the error in the code to run input through the LSTM layer.

PyTorch

output, (hn, cn) = lstm([1])

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

Ainput_tensor

Bx

Cinputs

Dinput

Attempts:

3 left

💡 Hint

Common Mistakes

Using undefined variable names.

Passing the wrong variable to the LSTM.

The variable input_tensor is the correct input tensor to pass to the LSTM.

4fill in blank

hard

Fill both blanks to create an LSTM with 2 layers and batch_first enabled.

PyTorch

lstm = nn.LSTM(input_size=10, hidden_size=20, num_layers=[1], batch_first=[2])

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

A2

BTrue

CFalse

D1

Attempts:

3 left

💡 Hint

Common Mistakes

Setting num_layers to 1 instead of 2.

Forgetting to set batch_first to True when input shape is batch first.

num_layers=2 creates two stacked LSTM layers. batch_first=True makes input shape (batch, seq, feature).

5fill in blank

hard

Fill all three blanks to extract the last hidden state from the LSTM output.

PyTorch

last_hidden = hn[[1], [2], [3]]

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

A-1

B0

C1

D2

Attempts:

3 left

💡 Hint

Common Mistakes

Using positive indices for the last layer instead of -1.

Mixing up batch and hidden unit indices.

hn[-1, 0, 1] accesses the last layer's hidden state for the first batch element and second hidden unit.