🤗 Large ML models in production with HuggingFace CTO Julien Chaumond from The MLOps Podcast

🤗 Large ML models in production with HuggingFace CTO Julien Chaumond

1 hour 19 minutes Posted Jul 4, 2021 at 8:07 am.

Guest intro

Origin of HuggingFace

Why the focus on NLP?

The success of the HuggingFace community

Reproducing models and scaling for the community

Enabling large models in production

How HuggingFace scales so many models

The biggest challenge HuggingFace solved in MLOps

How HuggingFace transitions from research to production

Using notebooks vs python modules

The most interesting topic in ML production

Fascinating ML research

Learning new things

Something that is true but most people disagree with

Tips to organize research teams

New features for accelerated inference

Most common use case of HuggingFace

Integrating search algorithms into transformer library

Integrating vision models

Long term business model

Automation and simplification of the process of building models

Support for real-time inference

Recommendations for the audience

0:00

1:19:11

Download MP3

Share at current time

Show notes

In this episode, I'm speaking with Julien Chaumond from 🤗 HuggingFace, about how they got started, getting large language models to production in millisecond inference times, and the CERN for machine learning.

Join our Discord community: https://discord.gg/tEYvqxwhah

---

Timestamps:

---

Relevant Links:

FastDS: https://github.com/DAGsHub/fds

BigScience: https://bigscience.huggingface.co

https://www.linkedin.com/company/dagshub/

https://www.linkedin.com/company/huggingface/

https://twitter.com/TheRealDAGsHub

https://twitter.com/huggingface

← See all 35 episodes of The MLOps Podcast