Google introduced its latest innovation, Gemma, which marks a completely new standard in the artificial intelligence industry. Essentially, Gemma is Google’s open source model for AI research. Developed by the brilliant minds of Google DeepMind and other teams, Gemma represents a significant breakthrough in bringing cutting-edge technology to AI developers and researchers around the world. This company, which takes its name from the Latin word meaning “precious stones”, represents the value that Google adds to the rapidly growing industry.

Google removed sensitive data from training sets to ensure model security

The Gemma family is available in two sizes, Gemma 2B and Gemma 7B, each designed to meet a variety of societal needs. Google has gone to great lengths to ensure these models are not only powerful but also responsible. The company ensures the safety of the model by extracting sensitive data from training sets and undergoing rigorous testing, including manual selection and automated competitive testing.

Gemma stands out with her versatility. Whether you’re working from a laptop or deploying applications to mobile, IoT or cloud platforms, Gemma is designed for seamless integration. Its compatibility with Vertex AI and Google Kubernetes Engine (GKE) simplifies deployment, and its partnership with Nvidia ensures Gemma runs optimally on both Google Cloud TPUs and Nvidia GPUs. These models not only match their larger counterparts, but outperform them in many cases. This level of performance opens up new opportunities for developers and delivers best-in-class performance that was previously unattainable.

Additionally, Google promotes the responsible use of AI by enabling commercial use and deployment in organizations of all sizes. With the launch of the Responsible Productive AI Toolkit, developers are putting resources at their fingertips to ensure Gemma is used ethically and efficiently.