Hugging Face
The open hub for models and datasets
WHAT IT IS
Hugging Face was founded in 2016, started as a chatbot app, then turned to distributing models and datasets. It maintains open-source libraries such as Transformers, Diffusers and Datasets, and gathers model repositories, datasets and online demos into one platform.
Why it matters
It turned the open distribution of models and datasets into a standard — unified formats, versions and interfaces — making downloading, fine-tuning and comparing models as easy as installing software, and becoming the de facto hub of the open-weight ecosystem.
Key milestones
- 2016
Founded
- 2018
Released the Transformers library
- 2021
Launched Spaces demos
Representative products
5Hugging Face Hub
2016The open clearing house for models and datasets
Transformers
2018One API to load and train pretrained models
Datasets
2020Load and stream datasets with one line of code
Spaces
2021Host a model demo as a web page people can open
Diffusers
2022A unified implementation and scheduler for diffusion models