How Gemma 3n can vary its memory footprint
Hey Cool Supporters!
As you probably already heard, Gemma 3n presented yesterday by Google can use different amounts of parameters during inference (from 2B to 5B) depending on the device it runs on. In this quick post I explain how it works.
Keep reading with a 7-day free trial
Subscribe to True Positive Weekly to keep reading this post and get 7 days of free access to the full post archives.