Google introduces new open-source AI models named Gemma

By Ovunc Kutlu

ISTANBUL (AA) - Google introduced its new generation of open-source artificial intelligence (AI) models on Wednesday, named Gemma, to help developers and researchers build AI responsibly.

"Gemma is a family of lightweight, state-of-the-art open models built from the same research and technology used to create the Gemini models," the company said in a statement.

"A new Responsible Generative AI Toolkit provides guidance and essential tools for creating safer AI applications with Gemma," it added.

Google said it is releasing model weights in two sizes, Gemma 2B and Gemma 7B, each with pre-trained and instruction-tuned variants that can run on laptops, workstations or Google Cloud.

It added that it has partnered with chip manufacturer Nvidia to optimize Gemma for Nvidia's graphics processing units, also known as GPUs.

Gemma comes less than one week after Google unveiled its new AI model, Gemini 1.5, formerly known as the generative AI chatbot Bard.

The company said Gemma will be available worldwide starting Wednesday.