Author(s): Krishan Walia

Originally published on Towards AI.

Just 2 GB of RAM is what it requires!

For decades, ‘powerful AI’ meant ‘big servers.’ Today, Google flips the script with Gemma-3n — an AI so compact it fits in 2 GB of memory yet outpaces larger models in speed🚀

Gemma 3n revolutionizes artificial intelligence by enabling it to function on edge and mobile devices, addressing user concerns about data privacy and the need for open-source solutions. The new model enhances performance without the necessity of cloud reliance, facilitating instantaneous language translation and offline AI applications. Furthermore, innovative technologies, including Per-Layer Embeddings, offer efficient use of memory, allowing complex AI models to operate seamlessly on devices with as little as 2 GB of RAM. This shift towards on-device processing not only ensures data safety but also paves the way for next-generation applications, effectively reimagining user interactions with AI through unprecedented speed and responsiveness.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI