Hugging Face in Action

Building with Hugging Face can mean much more than loading a model. In Hugging Face in Action, Wei-Meng Lee follows the work from discovering models and datasets to developing applications that use them—covering language, vision, agents, local data, and user interfaces along the way.

The book’s 13 chapters pair an introduction to the ecosystem with practical programming topics. For Python developers who know NumPy and Pandas, it offers a guided look at how the pieces can fit together in real AI projects.

Start with the Hugging Face ecosystem

Early chapters introduce the Transformers library, models, the Hub, and the Hugging Face approach to finding and using resources. The book then gets readers set up with the relevant Python tools before turning to transformer concepts, pipelines, and NLP tasks such as classification, generation, summarization, translation, and question answering.

Work across language, vision, and data

The scope extends beyond text. Separate chapters explore computer vision tasks—including object detection, image classification, segmentation, and video classification—and the discovery, preparation, tokenization, and visualization of datasets. Later, Lee covers fine-tuning pretrained models and working with multimodal models, connecting model concepts to practical workflows.

Build applications with LLMs and agents

For application development, the book introduces LangChain and LlamaIndex, including ways to connect language models with private data. It also looks at designing LangChain flows visually with Langflow, programming agents with tools and memory, and building web interfaces using Gradio. Further chapters explore locally running LLM applications with GPT4All, querying local data, and the Model Context Protocol.

A practical route from tools to projects

Rather than focusing on one model or use case, Hugging Face in Action surveys a connected toolkit for developing AI applications. Its broad coverage can help Python programmers understand how model libraries, datasets, application frameworks, and interfaces relate—and identify which areas to explore for their own projects.

Written for: Python programmers familiar with NumPy and Pandas. The publisher indicates that previous AI experience is not required.