Title: Training of Modern Large Language Models

Abstract: The talk will showcase the development of modern large language models such as the Falcon3 series of open models, with an emphasis on the specificities of training with knowledge distillation and up-scaling techniques, which allow significant pre-training cost reduction to build the models. We will also discuss models’ performances and capabilities by providing some insights on the main components to build large efficient language models.

Dates

March 15th, 2025

Abstract submission deadline

March 15th, 2025

Paper submission deadline

April 16th, 2025

Accept/Reject notification

May 21-23 ,2025

Netys Conference

Proceedings

The proceedings of NETYS 2025 are available at Springer’s LNCS Lecture Notes in Computer Science

Partners & Sponsors