Google Introduces Gemma: The AI Revolution
Google takes a turn towards the universalization of artificial intelligence (AI) with the launch of gema language model of Open Source of significant dimensions. Now, developers, researchers and AI lovers can explore a universe of possibilities without restrictions, invited to create the future of technology.
This initiative seeks to democratize access to artificial intelligence through the release of two open source large language models (LLM): Gemma 2B and Gemma 7B.
The Gemma models have been designed on the technical and infrastructure basis of Gemini, the largest and most powerful available today, allowing both Gemma 2B as 7B (the letter b represents ‘Billions’ of configuration parameters) achieve exceptional performance for its size, outperforming other open source models such as Flame-2 (Meta’s open source AI).
The key distinction between Gemma and Gemini lies in their approach: gem is primarily aimed at developers, while Gemini It is aimed at satisfying the needs of the final consumer.
Furthermore, these models can be run directly on a laptop or desktop computer from any developer, enabling widely accessible AI capabilities.
The most notable thing is that, despite its size, Gemma outperforms considerably larger models in key benchmarks maintaining high standards of safety and responsibility in its results.
Main features
✓ Open source
Developers around the world can freely access and modify the model codeenabling greater transparency, collaboration and innovation in the field of AI.
✓ Scalability
The Gemma models can run on a wide range of devicesfrom powerful cloud servers to low-power mobile devices, making them accessible for a wide variety of projects.
✓ Efficiency
The models are optimized for efficient use of power and memorywhich makes them more sustainable and economical.
✓ Responsibility
Google has made available a responsible AI toolkit which helps developers create ethical and secure applications.
Tool kit includes:
- Advanced security rating: Provides a unique methodology that allows building robust security classifiers with few examples, ensuring the protection of applications and users.
- Intuitive debugging tools– With this model debugging tool, you can easily investigate Gemma’s behavior, identify potential problems, and quickly fix them.
- Good practice guide– Access best practices in developing and implementing large language models.
Support for research and development:
Gemma is designed for the community of developers and researchers that drive innovation in artificial intelligence.
- Free access on Kaggle: start your project with Gemma at no cost using the free access on the platform kaggle.
- Free tier for Colab notebooks– Use resources without restrictions with Gemma’s free tier on Colab Notebooks.
- $300 in Google Cloud credits– As a new Google Cloud user, you will receive $300 in credits to start your projects with Gemma.
- Credits for researchers: Researchers can request credits of up to $500,000 in Google Cloud to further accelerate his investigations with Gemma.