Transformer architecture in NLP - Model Pipeline Trace

Model Pipeline - Transformer architecture

The Transformer architecture processes text data by converting words into numbers, then learning relationships between words using attention. It trains to predict the next word or classify text, improving accuracy over time.

Data Flow - 7 Stages

1Input Text

1 sentence x variable length→Raw text input→1 sentence x variable length

"The cat sat on the mat."

↓

2Tokenization

1 sentence x variable length→Split sentence into tokens (words or subwords)→1 sentence x 6 tokens

["The", "cat", "sat", "on", "the", "mat"]

↓

3Embedding

1 sentence x 6 tokens→Convert tokens to vectors of size 512→1 sentence x 6 tokens x 512 features

[[0.1, 0.3, ..., 0.2], ..., [0.05, 0.4, ..., 0.1]]

↓

4Positional Encoding

1 sentence x 6 tokens x 512 features→Add position info to embeddings→1 sentence x 6 tokens x 512 features

Embedding vectors with added position signals

↓

5Multi-Head Self-Attention

1 sentence x 6 tokens x 512 features→Calculate attention scores and weighted sums→1 sentence x 6 tokens x 512 features

Attention output vectors showing word relationships

↓

6Feed-Forward Network

1 sentence x 6 tokens x 512 features→Apply two linear layers with ReLU activation→1 sentence x 6 tokens x 512 features

Processed feature vectors for each token

↓

7Output Layer

1 sentence x 6 tokens x 512 features→Project to vocabulary size for prediction→1 sentence x 6 tokens x 10000 classes

Probabilities for each word in vocabulary

Training Trace - Epoch by Epoch

Loss
5.2 |**************
4.0 |**********
2.8 |*******
1.6 |****
0.4 |*
    +----------------
     1  3  5  7  10 Epochs

Epoch	Loss ↓	Accuracy ↑	Observation
1	5.2	0.12	Model starts with high loss and low accuracy
2	3.8	0.28	Loss decreases, accuracy improves as model learns
3	2.7	0.45	Model captures basic word relationships
4	1.9	0.60	Attention mechanism helps improve predictions
5	1.3	0.72	Model learns complex context and syntax
6	0.9	0.81	Loss steadily decreases, accuracy rises
7	0.7	0.86	Model converges with good performance
8	0.6	0.89	Fine tuning improves accuracy further
9	0.55	0.91	Model predictions become more confident
10	0.50	0.93	Training converges with low loss and high accuracy

Prediction Trace - 6 Layers

Layer 1: Tokenization

Layer 2: Embedding

Layer 3: Positional Encoding

Layer 4: Multi-Head Self-Attention

Layer 5: Feed-Forward Network

Layer 6: Output Layer

Model Quiz - 3 Questions

What does the positional encoding add to the token embeddings?

A Random noise to improve generalization

B Information about the order of words

C Labels for each token's part of speech

D The final prediction probabilities

Key Insight

The Transformer architecture uses attention to understand relationships between words regardless of their position. This allows it to learn complex language patterns efficiently, shown by decreasing loss and increasing accuracy during training.

Practice

1. What is the main purpose of the self-attention mechanism in a Transformer model?

easy

A. To increase the number of layers in the model

B. To reduce the size of the input data

C. To convert words into numbers

D. To let the model focus on different words in the sentence at the same time

Solution

1. Step 1: Understand self-attention role

   Self-attention helps the model look at all words together and decide which words are important for each word.

2. Step 2: Match purpose with options

   To let the model focus on different words in the sentence at the same time correctly describes this as focusing on different words simultaneously, unlike other options which describe unrelated tasks.

3. Final Answer:

   To let the model focus on different words in the sentence at the same time -> Option D

4. Quick Check:

   Self-attention = focus on words together [OK]

Hint: Self-attention means focusing on all words at once [OK]

Common Mistakes:

• Thinking self-attention reduces input size
• Confusing self-attention with embedding
• Assuming it increases model layers

2. Which of the following is the correct way to describe the Transformer architecture components?

easy

A. It has encoder and decoder parts

B. It has only an encoder part

C. It uses only convolutional layers

D. It uses recurrent neural networks

Solution

1. Step 1: Recall Transformer structure

   Transformers have two main parts: encoder to process input and decoder to generate output.

2. Step 2: Compare options with structure

   It has encoder and decoder parts correctly states the presence of both encoder and decoder; others mention incorrect or unrelated components.

3. Final Answer:

   It has encoder and decoder parts -> Option A

4. Quick Check:

   Transformer = encoder + decoder [OK]

Hint: Remember: Transformer = encoder + decoder [OK]

Common Mistakes:

• Thinking Transformer has only encoder
• Confusing Transformer with CNN or RNN
• Ignoring decoder role

3. Consider this simplified Transformer encoder code snippet in Python using PyTorch:

import torch
from torch import nn

class SimpleEncoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.attention = nn.MultiheadAttention(embed_dim=4, num_heads=2)
    def forward(self, x):
        attn_output, _ = self.attention(x, x, x)
        return attn_output

x = torch.rand(5, 3, 4)  # sequence length=5, batch=3, embed=4
model = SimpleEncoder()
output = model(x)
print(output.shape)

What will be the printed output shape?

medium

A. torch.Size([3, 5, 4])

B. torch.Size([5, 3, 4])

C. torch.Size([5, 4, 3])

D. torch.Size([3, 4, 5])

Solution

1. Step 1: Understand input shape and MultiheadAttention

   Input shape is (sequence length=5, batch=3, embedding=4). PyTorch MultiheadAttention expects (seq_len, batch, embed).

2. Step 2: Output shape matches input shape

   MultiheadAttention returns output with the same shape as input: (5, 3, 4).

3. Final Answer:

   torch.Size([5, 3, 4]) -> Option B

4. Quick Check:

   Output shape = input shape for MultiheadAttention [OK]

Hint: MultiheadAttention output shape matches input shape [OK]

Common Mistakes:

• Mixing batch and sequence dimensions
• Assuming output shape changes embedding size
• Confusing PyTorch input format

4. You have this Transformer decoder code snippet that throws an error:

import torch
from torch import nn

class SimpleDecoder(nn.Module):
    def __init__(self):
        super().__init__()
        self.attention = nn.MultiheadAttention(embed_dim=8, num_heads=4)
    def forward(self, tgt, memory):
        attn_output, _ = self.attention(tgt, memory, memory)
        return attn_output

tgt = torch.rand(10, 2, 8)  # target seq len=10, batch=2, embed=8
memory = torch.rand(5, 3, 8)  # memory seq len=5, batch=3, embed=8
model = SimpleDecoder()
output = model(tgt, memory)
print(output.shape)

What is the likely cause of the error?

medium

A. Sequence length mismatch between tgt and memory

B. Mismatch in embedding dimensions between tgt and memory

C. Batch size mismatch between tgt and memory

D. Number of attention heads is too high

Solution

1. Step 1: Check shapes of tgt and memory

   tgt=(10,2,8), memory=(5,3,8). Both have embedding size 8, sequence lengths differ (10 vs 5, allowed), but batch sizes differ (2 vs 3).

2. Step 2: Identify batch size mismatch

   Batch size mismatch between tgt (batch=2) and memory (batch=3) causes the RuntimeError in MultiheadAttention.

3. Step 3: Re-examine options carefully

   Embedding sizes match, sequence length mismatch is allowed, number of heads is valid. Batch size mismatch is most common error in such cases.

4. Final Answer:

   Batch size mismatch between tgt and memory -> Option C

5. Quick Check:

   Batch sizes must match for attention [OK]

Hint: Check batch sizes first when attention errors occur [OK]

Common Mistakes:

• Assuming sequence length must match
• Blaming embedding size mismatch incorrectly
• Thinking number of heads causes shape errors

5. You want to build a Transformer model for text summarization. Which combination of components is best suited for this task?

hard

A. Encoder-decoder, because summarization needs understanding input and generating output

B. Decoder only, because summarization is text generation

C. Neither encoder nor decoder, use RNN instead

D. Encoder only, because summarization needs understanding input only

Solution

1. Step 1: Understand summarization task

   Summarization requires reading input text (encoding) and producing a shorter text (decoding).

2. Step 2: Match task with Transformer parts

   Encoder-decoder architecture fits best as encoder understands input and decoder generates summary output.

3. Final Answer:

   Encoder-decoder, because summarization needs understanding input and generating output -> Option A

4. Quick Check:

   Summarization = encoder + decoder [OK]

Hint: Summarization needs both understanding and generating text [OK]

Common Mistakes:

• Choosing encoder only for generation tasks
• Choosing decoder only ignoring input understanding
• Ignoring Transformer benefits and choosing RNN