Google has introduced Gemma, a new set of open AI models designed to support developers and researchers in creating AI responsibly. Gemma models, which include the Gemma 2B and Gemma 7B, are built on the same technology as Google's Gemini models and are meant to be lightweight yet state-of-the-art. These models come in both pre-trained and instruction-tuned variants, accompanied by a Responsible Generative AI Toolkit for safer AI applications. Gemma is compatible with multiple frameworks like JAX, PyTorch, and TensorFlow, and can be run on various devices including laptops and cloud platforms. Google emphasizes safety and responsible use, with tools and methodologies for reducing risks and ensuring the models adhere to Google's AI Principles.

Gemma: Introducing new state-of-the-art open models