Evaluating pretrained text generation model

The PyBooks team has used a pre-trained GPT-2 model that you experimented to generate a text based on a given prompt. Now, they want to evaluate the quality of this generated text. To achieve this, they have tasked you to evaluate generated text using a reference text.

BLEUScore, ROUGEScore have been loaded for you.

This exercise is part of the course

Deep Learning for Text with PyTorch

Exercise instructions

Begin by initializing the two metrics (BLEU and ROUGE) provided from torchmetrics.text.
Use these initialized metrics to calculate the scores between the generated text and the reference text.
Display the calculated BLEU and ROUGE scores.

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

reference_text = "Once upon a time, there was a little girl who lived in a village near the forest."
generated_text = "Once upon a time, the world was a place of great beauty and great danger. The world of the gods was the place where the great gods were born, and where they were to live."

# Initialize BLEU and ROUGE scorers
bleu = ____()
rouge = ____()

# Calculate the BLEU and ROUGE scores
bleu_score = bleu([____], [[reference_text]])
rouge_score = rouge([generated_text], [[____]])

# Print the BLEU and ROUGE scores
print("BLEU Score:", bleu_score.____())
print("ROUGE Score:", rouge_score)

Edit and Run Code

This exercise is part of the course

Deep Learning for Text with PyTorch

AdvancedSkill Level

4.8+

Start Course for Free

This chapter introduces you to deep learning for text and its applications. Learn how to use PyTorch for text processing and get hands-on experience with techniques such as tokenization, stemming, stopword removal, and more. Understand the importance of encoding text data and implement encoding techniques using PyTorch. Finally, consolidate your knowledge by building a text processing pipeline combining these techniques.

Exercise 1: Introduction to preprocessing for text Exercise 2: Word frequency analysis Exercise 3: Preprocessing text Exercise 4: Encoding text data Exercise 5: One-hot encoded book titles Exercise 6: Bag-of-words for book titles Exercise 7: Applying TF-IDF to book descriptions Exercise 8: Introduction to building a text processing pipeline Exercise 9: Shakespearean language preprocessing pipeline Exercise 10: Shakespearean language encoder

Explore text classification and its role in Natural Language Processing (NLP). Apply your skills to implement word embeddings and develop both Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) for text classification using PyTorch, and understand how to evaluate your models using suitable metrics.

Exercise 1: Overview of Text Classification Exercise 2: Embedding in PyTorch Exercise 3: Categorizing text classification tasks Exercise 4: Convolutional neural networks for text classification Exercise 5: Build a CNN model for text Exercise 6: Train a CNN model for text Exercise 7: Testing the Sentiment Analysis CNN Model Exercise 8: Recurrent neural networks for text classification Exercise 9: Building an RNN model for text Exercise 10: Building an LSTM model for text Exercise 11: Building a GRU model for text Exercise 12: Evaluation metrics for text classification Exercise 13: Evaluating RNN classification models Exercise 14: Evaluating the model's performance Exercise 15: Comparing models

Venture into the exciting world of text generation and its applications in NLP. Understand how to leverage Recurrent Neural Networks (RNNs), Generative Adversarial Networks (GANs), and pre-trained models for text generation tasks using PyTorch. Alongside, you'll learn to evaluate the performance of your models using relevant metrics.

Exercise 1: Introduction to text generation Exercise 2: Creating a RNN model for text generation Exercise 3: Text generation using RNN - Training and Generation Exercise 4: Generative adversarial networks for text generation Exercise 5: Building a generator and discriminator Exercise 6: Training a GAN model Exercise 7: Pre-trained models for text generation Exercise 8: Text completion with pre-trained GPT-2 models Exercise 9: Language translation with pretrained PyTorch model Exercise 10: Evaluation metrics for text generation Exercise 11: Evaluating pretrained text generation model

Current Exercise

Exercise 12: Understanding text generation metrics

Understand the concept of transfer learning and its application in text classification. Explore Transformers, their architecture, and how to use them for text classification and generation tasks. You will also delve into attention mechanisms and their role in text processing. Finally, understand the potential impacts of adversarial attacks on text classification models and learn how to protect your models.

Exercise 1: Transfer learning for text classification Exercise 2: Transfer learning using BERT Exercise 3: Evaluating the BERT model Exercise 4: Transformers for text processing Exercise 5: Creating a transformer model Exercise 6: Training and testing the Transformer model Exercise 7: Attention mechanisms for text processing Exercise 8: Creating a RNN model with attention Exercise 9: Training and testing the RNN model with attention Exercise 10: Adversarial attacks on text classification models Exercise 11: Adversarial attack classification Exercise 12: Safeguarding AI at PyBooks Exercise 13: Wrap-up