BERT fine-tuning for classification in NLP

Introduction

BERT fine-tuning helps a pre-trained language model learn to classify text into categories. It saves time and works well even with small data.

You want to sort emails into spam or not spam.

You need to detect the sentiment of movie reviews as positive or negative.

You want to classify news articles by topic like sports, politics, or tech.

You have a small dataset but want good text classification results.

You want to improve a chatbot's understanding of user intent.

Syntax

NLP

from transformers import BertForSequenceClassification, BertTokenizer
from torch.utils.data import DataLoader
import torch

# Load pre-trained BERT model and tokenizer
model = BertForSequenceClassification.from_pretrained('bert-base-uncased', num_labels=2)
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')

# Prepare data: tokenize texts
inputs = tokenizer(texts, padding=True, truncation=True, return_tensors='pt')
labels = torch.tensor(labels)

dataset = torch.utils.data.TensorDataset(inputs['input_ids'], inputs['attention_mask'], labels)
dataloader = DataLoader(dataset, batch_size=8)

# Training loop example
optimizer = torch.optim.Adam(model.parameters(), lr=5e-5)
model.train()
for epoch in range(3):
    for batch in dataloader:
        input_ids, attention_mask, labels = batch
        outputs = model(input_ids, attention_mask=attention_mask, labels=labels)
        loss = outputs.loss
        loss.backward()
        optimizer.step()
        optimizer.zero_grad()

Use BertForSequenceClassification for classification tasks.

Tokenize text with padding and truncation to fit BERT's input size.

Examples

Load BERT for a 3-class classification problem.

NLP

model = BertForSequenceClassification.from_pretrained('bert-base-uncased', num_labels=3)

Tokenize a single sentence with padding and truncation.

NLP

inputs = tokenizer(['Hello world!'], padding=True, truncation=True, return_tensors='pt')

Get loss and prediction scores (logits) from the model.

NLP

outputs = model(input_ids, attention_mask=attention_mask, labels=labels)
loss = outputs.loss
logits = outputs.logits

Sample Model

This code fine-tunes BERT on two example sentences for sentiment classification. It prints the loss and predicted classes after one training pass.

NLP

from transformers import BertForSequenceClassification, BertTokenizer
from torch.utils.data import DataLoader, TensorDataset
import torch

# Sample data
texts = ['I love this movie', 'This movie is bad']
labels = [1, 0]  # 1=positive, 0=negative

# Load model and tokenizer
model = BertForSequenceClassification.from_pretrained('bert-base-uncased', num_labels=2)
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')

# Tokenize
inputs = tokenizer(texts, padding=True, truncation=True, return_tensors='pt')
labels_tensor = torch.tensor(labels)

dataset = TensorDataset(inputs['input_ids'], inputs['attention_mask'], labels_tensor)
dataloader = DataLoader(dataset, batch_size=2)

# Optimizer
optimizer = torch.optim.Adam(model.parameters(), lr=5e-5)

# Training loop (1 epoch for demo)
model.train()
for batch in dataloader:
    input_ids, attention_mask, labels_batch = batch
    outputs = model(input_ids, attention_mask=attention_mask, labels=labels_batch)
    loss = outputs.loss
    logits = outputs.logits
    loss.backward()
    optimizer.step()
    optimizer.zero_grad()

# Evaluation
model.eval()
with torch.no_grad():
    outputs = model(input_ids, attention_mask=attention_mask)
    predictions = torch.argmax(outputs.logits, dim=1)

print(f'Loss after training: {loss.item():.4f}')
print(f'Predictions: {predictions.tolist()}')

Important Notes

Fine-tuning usually needs a GPU for faster training.

Use a small learning rate like 5e-5 to avoid breaking the pre-trained model.

More epochs and data improve accuracy but take longer.

Summary

BERT fine-tuning adapts a powerful language model to your classification task.

Tokenize text properly before feeding it to BERT.

Train with a small learning rate and check loss and predictions to see progress.

Practice

1. What is the main purpose of fine-tuning BERT for a classification task?

easy

A. To adapt BERT's knowledge to classify specific categories in your data

B. To train BERT from scratch on a large dataset

C. To reduce the size of the BERT model for faster inference

D. To convert text into images for classification

Solution

1. Step 1: Understand BERT's pretraining

   BERT is pretrained on general language tasks and needs adjustment for specific tasks like classification.

2. Step 2: Purpose of fine-tuning

   Fine-tuning adapts BERT's learned language understanding to classify categories in your dataset.

3. Final Answer:

   To adapt BERT's knowledge to classify specific categories in your data -> Option A

4. Quick Check:

   Fine-tuning = adapt BERT for classification [OK]

Hint: Fine-tuning means adjusting BERT for your task, not training from scratch [OK]

Common Mistakes:

• Thinking fine-tuning trains BERT from zero
• Confusing fine-tuning with model compression
• Assuming BERT outputs images

2. Which of the following is the correct way to tokenize text before feeding it to BERT in Python?

easy

A. tokens = text.split(' ')

B. tokens = tokenizer.encode_plus(text, return_tensors='pt')

C. tokens = tokenizer.tokenize(text)

D. tokens = text.lower()

Solution

1. Step 1: Identify proper BERT tokenization method

   BERT uses tokenizer.encode_plus to convert text into token IDs and attention masks.

2. Step 2: Compare options

   tokens = tokenizer.encode_plus(text, return_tensors='pt') uses encode_plus with return_tensors='pt' for PyTorch tensors, which is correct for BERT input.

3. Final Answer:

   tokens = tokenizer.encode_plus(text, return_tensors='pt') -> Option B

4. Quick Check:

   Use encode_plus for BERT tokenization [OK]

Hint: Use tokenizer.encode_plus or tokenizer() for BERT input [OK]

Common Mistakes:

• Using simple split instead of tokenizer
• Only tokenizing without encoding IDs
• Not returning tensors for model input

3. Given this code snippet for fine-tuning BERT, what will be the output of print(predictions.argmax(dim=1)) if the model predicts logits [[2.0, 1.0], [0.5, 1.5]] for two samples?

logits = torch.tensor([[2.0, 1.0], [0.5, 1.5]])
predictions = logits
print(predictions.argmax(dim=1))

medium

A. tensor([2, 1])

B. tensor([1, 0])

C. tensor([1, 1])

D. tensor([0, 1])

Solution

1. Step 1: Understand argmax(dim=1)

   Argmax along dim=1 finds the index of max value in each row (sample).

2. Step 2: Calculate argmax for each sample

   First row: max is 2.0 at index 0; second row: max is 1.5 at index 1.

3. Final Answer:

   tensor([0, 1]) -> Option D

4. Quick Check:

   Argmax per row = [0, 1] [OK]

Hint: Argmax dim=1 picks max index per sample row [OK]

Common Mistakes:

• Confusing dim=0 with dim=1
• Mixing up indices and values
• Expecting values instead of indices

4. You run this training loop snippet but get a runtime error: TypeError: forward() missing 1 required positional argument: 'labels'. What is the likely fix?

outputs = model(input_ids, attention_mask)
loss = outputs.loss
loss.backward()

medium

A. Pass labels to the model call: model(input_ids, attention_mask, labels=labels)

B. Remove loss.backward() call

C. Change input_ids to input_id

D. Call model with only input_ids

Solution

1. Step 1: Understand error cause

   The model expects labels to compute loss but they are missing in the call.

2. Step 2: Fix by passing labels

   Include labels argument in model call to get loss: model(input_ids, attention_mask, labels=labels).

3. Final Answer:

   Pass labels to the model call: model(input_ids, attention_mask, labels=labels) -> Option A

4. Quick Check:

   Missing labels argument causes loss error [OK]

Hint: Always pass labels to get loss during training [OK]

Common Mistakes:

• Ignoring the missing labels argument
• Removing backward call instead of fixing input
• Changing variable names incorrectly

5. You want to fine-tune BERT on a small dataset for sentiment classification. Which strategy helps avoid overfitting during training?

hard

A. Train BERT without tokenization to save time

B. Increase batch size to maximum and train longer

C. Use a small learning rate and add dropout layers

D. Remove the classification head and train only embeddings

Solution

1. Step 1: Identify overfitting risks

   Small datasets can cause the model to memorize instead of generalize.

2. Step 2: Apply regularization techniques

   Using a small learning rate and dropout helps the model learn smoothly and avoid overfitting.

3. Final Answer:

   Use a small learning rate and add dropout layers -> Option C

4. Quick Check:

   Small LR + dropout reduces overfitting [OK]

Hint: Small learning rate + dropout helps generalize on small data [OK]

Common Mistakes:

• Training longer without regularization
• Skipping tokenization
• Removing classification head incorrectly