Gemma: Google’s New Open AI Models for Developers
Google’s Rapid Advancements in Generative AI Development
In the ever-evolving world of generative AI development, Google continues to make strides with its latest innovations. Just months after unveiling the groundbreaking Gemini model, Google has now introduced the next generation in its lineup: Gemma. Developed by DeepMind and other divisions, Gemma represents Google’s commitment to advancing responsible AI development.
Gemma: A Closer Look at Google’s Latest AI Models
Google’s announcement of the Gemma models emphasizes their focus on responsible AI development, leveraging the same cutting-edge research and technology that powers Gemini. Available in two sizes, the Gemma 2B and the Gemma 7B, these models come with pre-trained and instruction-tuned variants. Google touts Gemma’s ability to outperform larger models on key benchmarks while maintaining strict standards for safe and responsible outputs. Additionally, developers can run Gemma models locally on their computers, enhancing accessibility and usability.
Gemma in the Cloud: Enhancing AI Capabilities for Customers
Beyond local deployment, Gemma models are also integrated into Google Cloud, offering customers the opportunity to leverage them in Vertex AI and the Google Kubernetes Engine (GKE). With over 130 available models on Google Cloud, including Gemini, Gemma further expands the range of AI tools accessible to users. New Google Cloud customers can benefit from $300 in free credits towards utilizing Gemma, while researchers can apply for up to $500,000 in funding to support their projects. Additionally, Gemma is available for free on Kaggle, providing an open platform for experimentation and collaboration.
Gemma vs. Gemini: Distinguishing Between Google’s AI Offerings
While Gemini caters to a broader audience with public-facing products, Gemma is specifically tailored for developers and researchers. While developers can still utilize APIs to interact with Gemini for their projects, Gemma serves as a standalone open model with alternative access options. Google’s deliberate naming scheme underscores the relationship between Gemini and Gemma, both drawing inspiration from Latin roots. Just as Gemini represents twins, Gemma translates to “precious stone,” aligning with the gemstone theme shared between the two models.
By introducing Gemma alongside Gemini, Google continues to drive innovation in the field of AI development, offering developers and researchers a diverse range of tools and resources to fuel their projects. Stay tuned for more updates on Google’s latest AI models and their impact on the ever-evolving AI landscape.