PyTorch at scale for translation and NLP
PyText is a research to production platform that Facebook has leveraged to quickly develop state-of-the-art NLP systems and deploy them to critical production use cases. Stef Nelson-Lindall explores several challenges with developing, training, and deploying real production systems with Torch and how to deal with them in NLP use cases. He outlines replicability of pretraining and training, training speed, and resiliency of models once they’ve been successfully trained. You’ll learn about some of the developer efficiency problems with PyText and how the company is addressing them.
What you'll learn
- Learn about how PyText has been leveraged to develop and deploy NLP systems
Stef Nelson-Lindall is a tech lead for PyText, an open source Facebook project for experimentation, training, and productionization of NLP models using PyTorch. He worked to use this project to deploy Transformer architecture models to real-time production systems at Facebook. Previously, he built NLP and dialog systems for products across Facebook, built CRM systems at Google, and helped build web-based medical image viewers at Vital Images. He holds bachelors degrees in math and computer science from the University of Minnesota.
Diversity and Inclusion Sponsor
Premier Exhibitor Plus
R & D and Innovation Track Sponsor
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
For media/analyst press inquires