AI Rewind: Google's Open "Access" Gemma Models

Meta talks the talk but Google's walking the walk

Mark Hinkle
February 24, 2024

In a move that underscores Google's commitment to the democratization of artificial intelligence (AI), the tech giant has unveiled Gemma, a suite of open models set to redefine the landscape of AI development and application.

Gemma: Google's Open Model

Gemma emerges from the same innovative crucible that brought us the Gemini models (Google Bard is now Gemini for those who missed the relabeling). Gemma, drawn from the Latin word for "precious stone," symbolizes the model's value and potential impact on the AI community. Available in two variants, Gemma 2B and Gemma 7B, these models are designed to be lightweight yet state-of-the-art, suitable for various applications from the get-go.

A Toolkit for Responsible AI

Google's launch of Gemma is accompanied by a Responsible Generative AI Toolkit, which underscores the company's commitment to the ethical development and deployment of AI technologies. This comprehensive resource provides guidance, best practices, and essential tools to ensure that applications built with Gemma are safe, reliable, and aligned with responsible AI principles.

Optimized for Performance and Accessibility

What sets Gemma apart is its ethical foundation, technical prowess, and accessibility. Google has ensured that Gemma models are optimized for performance across a wide range of hardware platforms, including NVIDIA GPUs and Google Cloud TPUs, making state-of-the-art AI capabilities more accessible than ever before. Furthermore, the terms of use for Gemma encourage responsible commercial usage and distribution, making it a valuable asset for organizations of all sizes.

Performance Benchmarks for Gemini Models

Google Gemma vs. Meta's LLaMA: A More Open Approach

The introduction of Gemma provides a stark contrast to Meta's LLaMA models, particularly in terms of licensing and accessibility. While both initiatives aim to push the boundaries of AI research and application, Google's Gemma is distinguished by its open, collaborative ethos. I will say that Gemma models are not licensed under an open source license. Still, the restrictions outlined in the usage policy are all aimed at public safety and don’t appear to have any anti-competitive conditions.

LLama, on the other hand, has a few conditions that restrict competition.

You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Llama 2 or derivative works thereof).
Additional Commercial Terms. If, on the Llama 2 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee’s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights.

While Llama’s approach is fair, Google has much more on ethos. This I think not only fosters innovation but also facilitates a more inclusive AI development landscape, enabling a broader spectrum of developers and researchers to contribute to and benefit from advanced AI technologies.

The Future of AI with Gemma

As Google continues to expand the Gemma model family, the potential for transformative AI applications grows exponentially. The company's vision for Gemma is about providing powerful AI tools and building a community around responsible AI development. With free access to Gemma through platforms like Kaggle, Hugging Face, and Google Colab, and substantial Google Cloud credits for research and development, Google is laying the groundwork for a future where AI innovation is open, accessible, and responsible.

Conclusion

Google's launch of Gemma marks a significant milestone in the journey towards a more open, inclusive, and responsible AI future. By providing state-of-the-art AI models and a comprehensive toolkit for responsible development, Google is not just contributing to the AI research community; it's reshaping the fabric of how AI technologies are developed, deployed, and shared. As we stand on the brink of this new era, the possibilities are as boundless as they are exciting. Gemma isn't just a precious stone in Google's crown; it's a beacon for the future of AI.