In the first of two Cerebras Developer Community Meetings held in July 2022, Senior Deep Learning Engineer, Sia Rezaei, provides a deep dive into our latest capability, easily training 1.3 – 20 billion parameter GPT-style models, such as GPT-J, on a single CS-2 machine. This is record-setting for large language model training on a single device.

Add comment

Your email address will not be published. Required fields are marked *

Categories

All Topics