Overview

Google, IBM & Meta Certificates – 40% Off

One plan covers every Professional Certificate on Coursera.

Get ready to put your Generative AI, Natural Language Processing (NLP), and PyTorch skills into action in this hands-on capstone project from IBM. During this course, you’ll solve a real-world text classification challenge by building an end-to-end NLP workflow, from raw text processing to model evaluation. You’ll design and implement complete pipelines, including text preprocessing, tokenization, vocabulary creation, and dataset preparation using PyTorch Dataset and DataLoader. You’ll train and compare RNN, LSTM, and Transformer models, and explore how each architecture processes language differently. Plus, you’ll fine-tune pretrained models using Hugging Face Transformers, applying techniques used in production-grade AI systems. By the end of the course, you’ll have a portfolio-worthy capstone project that showcases your ability to build, optimize, and evaluate NLP models using metrics such as accuracy and F1-score. Great for talking about in interviews. Enroll today to strengthen your Generative AI and NLP skills and showcase your expertise with this powerful, job-oriented capstone project.

Syllabus

Text Data Preparation and Preprocessing

In this module, you will explore how text data is prepared for natural language processing workflows in PyTorch. You will work with text-loading strategies, tokenization methods, vocabulary construction, and batching techniques to create model-ready inputs. Through readings, guided activities, and hands-on labs, you will examine how preprocessing choices affect downstream model development. Additionally, you will also practice analyzing data preparation challenges in a realistic NLP workflow.

Sequential Text Classification with RNNs and LSTMs

In this module, you’ll explore how sequential models support text classification tasks in PyTorch. You’ll examine recurrent neural networks (RNNs), long short-term memory (LSTM) models, and sentiment analysis workflows. The hands-on labs enable you to train, evaluate, and refine these architectures while examining how regularization and optimizer choices affect convergence, generalization, and overall model performance. Finally, you will also compare RNN and LSTM results to identify performance trade-offs and justify the most effective architecture for the task.

Transformer Model Development and Fine-Tuning

In this module, you will explore how transformer architectures support modern NLP workflows. You will examine self-attention, positional encoding, tokenization, and transfer learning as the foundation for transformer-based text classification. You’ll use PyTorch to work with core transformer components and fine-tune a pretrained model using the Hugging Face ecosystem. Finally, you will also interpret evaluation results, compare tuning outcomes, and justify fine-tuning decisions using performance evidence.

Final Project and Course Wrap-Up

In this module, you will complete a cumulative final project that integrates skills from across the specialization. You will submit your output in Jupyter notebooks that demonstrate your proficiency in PyTorch, neural network design, and NLP techniques. Finally, you will also consolidate your learning with a course wrap-up and assess your understanding with a final exam.