Transformers-for-NLP-and-Computer-Vision-3rd-Edition

Why we included this project

This repository pairs with Denis Rothman's handbook on transformer models and collects a large set of hands-on notebooks for learning how modern LLMs and vision models work in practice. The lessons cover the transformer architecture, fine-tuning BERT and pretraining RoBERTa, working with the OpenAI API including the newer o1 and o3 reasoning models, and building on Hugging Face and Google Vertex AI. Vision material includes vision transformers, CLIP, Stable Diffusion, and DALL-E 3, along with practical sections on retrieval-augmented generation to cut down hallucinations and on interpretability tools like BertViz, LIME, and SHAP. Notebooks launch directly from GitHub into Colab, Kaggle, Gradient, or SageMaker Studio Lab, so you can follow along without setting up a local environment. The author keeps adding bonus notebooks for newer releases such as DeepSeek-R1 and GPT-4o, which keeps the collection useful long after publication. For anyone teaching themselves the field, it works as a map through both the theory and the working code.