Glossary Entry

What is Hugging Face?

Hugging Face is a company and platform hosting open models and datasets, known for its Transformers library and Model Hub. See how its ecosystem fits together.

Written by Alex Merced

What is Hugging Face?

Introduction to Hugging Face

Before 2018, machine learning was fragmented and hard to enter. If a researcher published a new AI model in an academic paper, replicating the work was difficult. A developer often had to work through the math, write custom PyTorch or TensorFlow code, and spend weeks training the model just to see if it worked.

Hugging Face made this distribution problem much easier.

Originally founded in 2016 as a chatbot app for teenagers, Hugging Face pivoted to become what many call the “GitHub of machine learning.” It is an open-source platform and community with a large repository that standardized how many AI models are shared and deployed.

The Core Pillars of Hugging Face

Hugging Face’s position in the AI industry is built on two pillars: the Python library and the Model Hub.

1. The transformers Python Library

In 2018, Hugging Face released the open-source Python library that became transformers. It provided a unified, simple API for downloading and running pre-trained neural networks. Instead of writing a large amount of model-loading code, a developer could run a natural language processing model in three lines of code:

from transformers import pipeline
classifier = pipeline("sentiment-analysis")
classifier("I love this product!")

This library lowered the barrier to AI engineering, allowing developers without a research background to add pre-trained models to their applications.

2. The Model Hub

The Hugging Face Model Hub is a central repository for the open-source AI community, hosting more than a million pre-trained models. When a company like Meta releases a Llama model, or Mistral releases a new model, they commonly publish the weight files on Hugging Face. The Hub is searchable by task, language, and license. If a developer needs a model trained to translate English to German, they can search the Hub, find a community fine-tuned model, and download the weights.

Expanding the Ecosystem

As the AI industry grew beyond text processing, Hugging Face expanded its platform to cover more of the MLOps lifecycle.

  • Datasets: Hugging Face hosts open datasets used to train and evaluate models, along with a datasets library that standardizes how data scientists load them.
  • Spaces: A hosting platform that lets developers deploy their models as interactive web applications (using Gradio or Streamlit) so the public can try them without writing code.
  • Inference Endpoints: For enterprise customers who do not want to manage their own infrastructure, Hugging Face can deploy a model from the Hub to a dedicated, autoscaling API endpoint.

Conclusion

Hugging Face is a central part of the open-source AI ecosystem. By providing standard libraries that hide much of the implementation complexity, and by hosting the Hub where researchers and companies share their models, Hugging Face has sped up the pace of open AI development.

Learn More

To learn more about the Data Lakehouse, read The Lakehouse Built for Everyone by Alex Merced. You can find this and other books by Alex Merced at books.alexmerced.com.

Related Terms

What is MLOps?

MLOps is the practice of applying DevOps methods to machine learning: versioning code and data, automating training and deployment, and watching for drift.

What is Natural Language Processing (NLP)?

Natural language processing (NLP) is the field of AI that lets computers parse, understand, and generate human language, from sentiment to transformers.

What is Machine Learning?

Machine learning is a branch of AI where algorithms learn patterns from data instead of fixed rules. See its main types and how lakehouse tables feed training.

What is Data Lakehouse?

A data lakehouse is an architecture that adds warehouse features like ACID transactions and schema enforcement to data lake files via open formats like Iceberg.

What is Open-Source LLMs (Llama, Mistral)?

Open-source LLMs are language models, such as Meta's Llama and Mistral's models, whose weights are public to run, and why companies host them privately.

What is Generative Pre-trained Transformer (GPT)?

GPT is a family of autoregressive language models that predict the next token. Learn what generative, pre-trained, and transformer each mean in how GPT works.

← Back to All Terms

Work with Alex