Sylvain Gugger

Career

Sylvain Gugger was a Research Scientist at fast.ai before joining Hugging Face as a Research Engineer, where Gugger became one of the core maintainers of the Transformers library. Hugging Face's course page describes Gugger's research focus as making deep learning more accessible through techniques that let models train quickly on limited resources.²¹

Open-source work

At fast.ai, Gugger co-wrote with Jeremy Howard the 2020 paper describing fastai's layered API, and the two share copyright on the fastbook notebooks, which are used for an online course and form the basis of the book Deep Learning for Coders with fastai and PyTorch.⁵⁶²

Gugger is a co-author of the 2019 paper introducing the Transformers library. In April 2021, Gugger introduced Accelerate on the Hugging Face blog as a library for PyTorch users who want to keep full control of their training loops while running them on several GPUs, TPUs, or in mixed precision.³⁷⁴

Work and contributions

• Transformers Open-source code

  Co-author of the 2019 paper introducing Hugging Face's Transformers library (arXiv 1910.03771), and described on Hugging Face's LLM Course page as one of the library's core maintainers. The repository is released under the Apache 2.0 license.³²⁸

• Accelerate Open-source code

  Listed first among the authors in the citation entry of the Accelerate README (2022). The Apache-2.0 library lets people who write their own PyTorch training loops run them in distributed or mixed-precision setups.⁴

• Deep Learning for Coders with fastai and PyTorch Published research

  Co-wrote the book with Jeremy Howard, according to the author biographies on Hugging Face's LLM Course page; the fastbook repository holds the notebooks the book is based on, copyright Jeremy Howard and Sylvain Gugger from 2020.²⁶

• fastai: A Layered API for Deep Learning Published research

  Co-author, with Jeremy Howard, of the 2020 paper describing the design of the fastai deep learning library as a set of layered APIs.⁵

Timeline

1. Oct 9, 2019Co-authored the paper introducing Hugging Face's Transformers library.³
2. 2020Published the fastbook notebooks with Jeremy Howard as the basis of their deep learning book.⁶
3. Feb 11, 2020Released "fastai, A Layered API for Deep Learning" with Jeremy Howard on arXiv.⁵
4. Apr 16, 2021Introduced the Accelerate library on the Hugging Face blog.⁷

Documented roles

The sources do not confirm a current role; these are the roles they document.

• Research Engineer and one of the core maintainers of the Transformers library, Hugging Face²¹
• Research Scientist, fast.ai²¹

Companies in their history

• Hugging Face 34 people in the wiki

Related people

• Clément DelangueConnected to Hugging Face
• Stas BekmanConnected to Hugging Face
• Thomas WolfConnected to Hugging Face
• Sasha RushConnected to Hugging Face
• Lewis TunstallConnected to Hugging Face
• Mu LiML frameworks and libraries · Teaching and writing
• Brian Van EssenTraining systems · ML frameworks and libraries
• Pieter NoordhuisTraining systems · ML frameworks and libraries
• Martín AbadiML frameworks and libraries · Training systems
• Chip HuyenTeaching and writing · ML frameworks and libraries