Positional encoding in PyTorch - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of Positional Encoding Tensor Shape

What is the shape of the positional encoding tensor generated by the following PyTorch code snippet?

PyTorch

import torch
import math

def positional_encoding(seq_len, d_model):
    pe = torch.zeros(seq_len, d_model)
    position = torch.arange(0, seq_len, dtype=torch.float).unsqueeze(1)
    div_term = torch.exp(torch.arange(0, d_model, 2).float() * (-math.log(10000.0) / d_model))
    pe[:, 0::2] = torch.sin(position * div_term)
    pe[:, 1::2] = torch.cos(position * div_term)
    return pe

pe_tensor = positional_encoding(50, 512)
print(pe_tensor.shape)

A (50, 512)

B (512, 50)

C (50, 256)

D (512, 50, 2)

💡 Hint

Think about how the positional encoding is created for each position and each dimension.

The positional encoding tensor has shape (sequence length, model dimension). Here, sequence length is 50 and model dimension is 512, so the shape is (50, 512).

🧠 Conceptual

intermediate

Purpose of Positional Encoding in Transformers

Why do transformer models use positional encoding?

A To normalize the input data before feeding it to the model.

B To add information about the order of tokens since transformers have no built-in sequence order awareness.

C To randomly shuffle the input tokens to improve generalization.

D To reduce the size of the input embeddings for faster computation.

💡 Hint

Transformers process all tokens simultaneously without sequence order. How do they know token positions?

Transformers do not have recurrence or convolution, so they cannot inherently understand token order. Positional encoding adds unique position information to each token embedding.

❓ Hyperparameter

advanced

Effect of Changing Model Dimension in Positional Encoding

If you increase the model dimension (d_model) in the positional encoding function, what is the expected effect on the positional encoding vectors?

A The positional encoding vectors will become sparse with many zeros.

B The positional encoding vectors will become shorter, losing positional information.

C The positional encoding vectors will have more dimensions, allowing finer granularity in encoding positions.

D The positional encoding vectors will remain the same size but with different values.

💡 Hint

Consider what d_model controls in the positional encoding tensor shape.

Increasing d_model increases the number of dimensions in the positional encoding, allowing the model to encode position information with more detail.

🔧 Debug

advanced

Identify the Error in Positional Encoding Code

What error will the following PyTorch code raise when executed?

PyTorch

import torch
import math

def positional_encoding(seq_len, d_model):
    pe = torch.zeros(seq_len, d_model)
    position = torch.arange(0, seq_len).float().unsqueeze(1)
    div_term = torch.exp(torch.arange(0, d_model, 2) * (-math.log(10000.0) / d_model))
    pe[:, 0::2] = torch.sin(position * div_term)
    pe[:, 1::2] = torch.cos(position * div_term)
    return pe

pe_tensor = positional_encoding(50, 512)

A RuntimeError due to shape mismatch in assignment to pe[:, 0::2]

B SyntaxError due to missing colon in function definition

C TypeError because torch.arange returns a list, not a tensor

D No error, code runs successfully

💡 Hint

Check the data types and shapes of tensors used in multiplication and assignment.

torch.arange returns a tensor of integers by default. Multiplying integer tensor with float tensor without casting causes shape mismatch in assignment.

❓ Model Choice

expert

Choosing Positional Encoding Type for a Transformer Model

You want to build a transformer model for a task with very long sequences (e.g., 10,000 tokens). Which positional encoding approach is best to handle this scenario?

A Use learned positional embeddings with fixed maximum length of 512 tokens.

B Use random positional embeddings initialized at each training batch.

C Use no positional encoding and rely on attention alone.

D Use sinusoidal positional encoding which can generalize to longer sequences.

💡 Hint

Consider which encoding can extrapolate beyond training sequence lengths.

Sinusoidal positional encoding uses fixed mathematical functions that can generalize to any sequence length, unlike learned embeddings limited by training length.

Practice

1. What is the main purpose of positional encoding in transformer models?

easy

A. To increase the number of layers in the model

B. To reduce the size of the input data

C. To add information about the order of tokens in a sequence

D. To normalize the input embeddings

Solution

1. Step 1: Understand the role of positional encoding

   Positional encoding adds information about the position of each token in the input sequence, which helps the model understand order. The other options relate to unrelated tasks such as increasing the number of layers, reducing input size, or normalizing embeddings.

2. Final Answer:

   To add information about the order of tokens in a sequence -> Option C

3. Quick Check:

   Positional encoding = order info [OK]

Hint: Positional encoding = position info added to embeddings [OK]

Common Mistakes:

• Thinking positional encoding reduces data size
• Confusing positional encoding with normalization
• Assuming it increases model layers

2. Which of the following is the correct way to compute the positional encoding for position pos and dimension i using sine and cosine functions in PyTorch?

easy

A. pos_enc[pos, i] = sin(pos * 10000^(i/d_model))

B. pos_enc[pos, 2*i] = sin(pos / 10000^(2*i/d_model)); pos_enc[pos, 2*i+1] = cos(pos / 10000^(2*i/d_model))

C. pos_enc[pos, i] = cos(pos / 10000^(i/d_model))

D. pos_enc[pos, 2*i] = cos(pos / 10000^(2*i/d_model)); pos_enc[pos, 2*i+1] = sin(pos / 10000^(2*i/d_model))

Solution

1. Step 1: Recall the positional encoding formula

   Positional encoding uses sine for even indices and cosine for odd indices with the formula: pos_enc[pos, 2*i] = sin(pos / 10000^(2*i/d_model)); pos_enc[pos, 2*i+1] = cos(pos / 10000^(2*i/d_model)).

2. Final Answer:

   pos_enc[pos, 2*i] = sin(pos / 10000^(2*i/d_model)); pos_enc[pos, 2*i+1] = cos(pos / 10000^(2*i/d_model)) -> Option B

3. Quick Check:

   Sine even, cosine odd = pos_enc[pos, 2*i] = sin(pos / 10000^(2*i/d_model)); pos_enc[pos, 2*i+1] = cos(pos / 10000^(2*i/d_model)) [OK]

Hint: Sine for even dims, cosine for odd dims in positional encoding [OK]

Common Mistakes:

• Swapping sine and cosine roles
• Using multiplication instead of division in exponent
• Applying sine or cosine to all dimensions only

3. Given the following PyTorch code snippet for positional encoding, what is the shape of the output tensor pos_encoding?

import torch
import math

def get_positional_encoding(seq_len, d_model):
    pos_encoding = torch.zeros(seq_len, d_model)
    for pos in range(seq_len):
        for i in range(0, d_model, 2):
            pos_encoding[pos, i] = math.sin(pos / (10000 ** (i / d_model)))
            if i + 1 < d_model:
                pos_encoding[pos, i + 1] = math.cos(pos / (10000 ** (i / d_model)))
    return pos_encoding

pos_encoding = get_positional_encoding(50, 512)

medium

A. [50, 512]

B. [512, 50]

C. [50, 256]

D. [512, 512]

Solution

1. Step 1: Analyze function input parameters and output shape

   The function takes seq_len=50 and d_model=512, creating a tensor of zeros with shape (50, 512). The loops fill values in this tensor without changing its shape, so output shape remains (50, 512).

2. Final Answer:

   [50, 512] -> Option A

3. Quick Check:

   Output shape = (seq_len, d_model) = [50, 512] [OK]

Hint: Output shape matches (sequence length, embedding dim) [OK]

Common Mistakes:

• Confusing dimensions order
• Thinking output is half the embedding size
• Assuming square matrix output

4. Identify the error in this PyTorch positional encoding code and select the fix:

import torch
import math

def positional_encoding(seq_len, d_model):
    pos_enc = torch.zeros(seq_len, d_model)
    for pos in range(seq_len):
        for i in range(d_model):
            pos_enc[pos, i] = math.sin(pos / 10000 ** (i / d_model))
    return pos_enc

pos_enc = positional_encoding(10, 16)

medium

A. Use sine for even indices and cosine for odd indices instead of sine only

B. Change math.sin to torch.sin for GPU compatibility

C. Initialize pos_enc with ones instead of zeros

D. Remove the inner loop and use vectorized operations

Solution

1. Step 1: Review positional encoding formula and identify fix

   Positional encoding uses sine for even dimensions and cosine for odd dimensions, but code uses sine only. Use sine for even indices and cosine for odd indices instead of sine only correctly fixes this.

2. Final Answer:

   Use sine for even indices and cosine for odd indices instead of sine only -> Option A

3. Quick Check:

   Sine and cosine alternate = Use sine for even indices and cosine for odd indices instead of sine only [OK]

Hint: Positional encoding alternates sine and cosine per dimension [OK]

Common Mistakes:

• Using sine only for all dimensions
• Ignoring cosine for odd indices
• Thinking initialization value causes error

5. You want to add positional encoding to a batch of input embeddings with shape (batch_size, seq_len, d_model) in PyTorch. Which approach correctly applies positional encoding to each sequence in the batch?

hard

A. Create a positional encoding tensor of shape (batch_size, d_model) and multiply element-wise with input embeddings

B. Create a positional encoding tensor of shape (seq_len,) and add it to input embeddings directly

C. Create a positional encoding tensor of shape (batch_size, seq_len, 1) and concatenate it to input embeddings

D. Create a positional encoding tensor of shape (seq_len, d_model) and add it to input embeddings using broadcasting

Solution

1. Step 1: Understand input and positional encoding shapes

   Input embeddings have shape (batch_size, seq_len, d_model). Positional encoding is per position and dimension, so shape (seq_len, d_model) fits. Adding it broadcasts over batch dimension correctly.

2. Final Answer:

   Create a positional encoding tensor of shape (seq_len, d_model) and add it to input embeddings using broadcasting -> Option D

3. Quick Check:

   Broadcast positional encoding over batch = Create a positional encoding tensor of shape (seq_len, d_model) and add it to input embeddings using broadcasting [OK]

Hint: Broadcast positional encoding (seq_len, d_model) over batch [OK]

Common Mistakes:

• Using positional encoding without seq_len dimension
• Multiplying instead of adding positional encoding
• Concatenating positional encoding incorrectly