Google has introduced Gemma, its latest innovation in the AI sector, representing a significant advancement in AI research. Developed by Google DeepMind and other teams, Gemma offers state-of-the-art technology openly available to developers and researchers worldwide. Named after the Latin word for “precious stone,” Gemma embodies Google’s commitment to the rapidly-growing AI field.
The Gemma family includes two models, Gemma 2B and Gemma 7B, tailored to meet diverse community needs. Google has prioritized model safety by removing sensitive data from training sets and undergoing rigorous testing, ensuring responsible AI development.
Gemma’s versatility stands out, seamlessly integrating across various platforms, including mobile, IoT, and cloud, with compatibility with Vertex AI and Google Kubernetes Engine (GKE). Partnering with Nvidia optimizes Gemma’s performance on Google Cloud TPUs and Nvidia GPUs, offering best-in-class efficiency and new possibilities for developers.
Furthermore, Google promotes responsible AI usage, allowing commercial distribution and providing resources like the Responsible Generative AI Toolkit to ensure ethical and effective Gemma usage across organizations of all sizes.
