Positional encoding in PyTorch - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is positional encoding in the context of machine learning models like Transformers?

Positional encoding is a way to add information about the order or position of elements in a sequence, so the model knows where each element is located since models like Transformers do not process data in order by default.

beginner

Why do Transformers need positional encoding?

Transformers process all input tokens simultaneously without any inherent order. Positional encoding helps the model understand the sequence order, which is important for tasks like language understanding.

intermediate

How is sinusoidal positional encoding calculated?

It uses sine and cosine functions of different frequencies to create unique position vectors. For position pos and dimension i: PE(pos, 2i) = sin(pos / 10000^(2i/d_model)) and PE(pos, 2i+1) = cos(pos / 10000^(2i/d_model)). This helps the model learn relative positions.

intermediate

What is the shape of the positional encoding tensor for a batch of sequences?

The positional encoding tensor usually has shape (sequence_length, embedding_dimension). When added to input embeddings, it matches the input shape (batch_size, sequence_length, embedding_dimension) by broadcasting over the batch.

beginner

How does adding positional encoding affect the input embeddings in a Transformer?

Positional encoding is added element-wise to the input embeddings. This combined input carries both the token meaning and its position, allowing the Transformer to use position information during training and prediction.

Why can't Transformers rely on sequence order without positional encoding?

A Because they ignore input data

B Because they only work with images

C Because they use recurrent layers

D Because they process all tokens simultaneously without order

Transformers process tokens in parallel, so they need positional encoding to know the order of tokens.

Which functions are used in sinusoidal positional encoding?

A Sine and cosine

B Linear and quadratic

C Tangent and cotangent

D Exponential and logarithm

Sinusoidal positional encoding uses sine and cosine functions to encode positions.

What does the positional encoding vector depend on?

A The batch size

B Only the token value

C The position in the sequence and embedding dimension

D The model's output

Positional encoding depends on the position index and the embedding dimension to create unique vectors.

How is positional encoding combined with input embeddings?

A By adding them element-wise

B By multiplying them

C By concatenating them

D By ignoring positional encoding

Positional encoding is added element-wise to input embeddings to include position information.

What is the main benefit of sinusoidal positional encoding?

A It speeds up training

B It allows the model to learn relative positions

C It reduces model size

D It removes the need for embeddings

Sinusoidal encoding helps the model understand relative positions between tokens.

Explain in your own words why positional encoding is important for Transformer models.

Think about how Transformers see all words at once.

Describe how sinusoidal positional encoding is calculated and why sine and cosine functions are used.

Focus on the math functions and their role in encoding position.

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