[Solved] Examine the following PyTorch GRU usage:gru = nn.GRU(input_size=20, hidden_size=40) input_tensor = torch.... | NLP

NLP - Sequence Models for NLP

Examine the following PyTorch GRU usage:

gru = nn.GRU(input_size=20, hidden_size=40)
input_tensor = torch.randn(10, 5, 20)
output, hidden = gru(input_tensor, hidden=None)

What is the issue with this code?

AThe input tensor shape does not match the expected shape when batch_first=False

BThe hidden_size must be equal to input_size

CThe GRU layer requires batch_first=True for this input

DThe hidden state should not be passed as None explicitly

Step-by-Step Solution

Solution:

Step 1: Check default batch_first value
By default, batch_first=False, so input shape should be (seq_len, batch, input_size).
Step 2: Compare input shape
Input tensor shape is (10, 5, 20) which is (seq_len=10, batch=5, input_size=20) - this matches default.
Step 3: Check hidden argument
Passing hidden=None is not necessary; if omitted, GRU initializes hidden state to zeros automatically.
Step 4: Identify the problem
Explicitly passing hidden=None is redundant and can cause confusion.
Final Answer:
The hidden state should not be passed as None explicitly -> Option D
Quick Check:
GRU initializes hidden state automatically if not provided [OK]

Quick Trick: Do not pass hidden=None explicitly; let GRU handle initialization [OK]

Common Mistakes:

MISTAKES

Master "Sequence Models for NLP" in NLP

9 interactive learning modes - each teaches the same concept differently

More NLP Quizzes

Examine the following PyTorch GRU usage: