Google released an open AI model called Gemwhich he claims is created using the same research and technology used to build his Gemini AI Models. The company says Gemma is its contribution to the open community and aims to help developers “build AI responsibly.” In this capacity, he also introduced the Responsible Generative AI Toolkit alongside Gemma. It contains a debugging tool, as well as a best practice guide for AI development based on Google's experience.
The company has made Gemma available in two different sizes – Gemma 2B and Gemma 7B – both of which come with pre-trained and instruction-optimized variants and are both lightweight enough to run directly on the laptop or desktop computer of a developer. Google claims that Gemma outperforms much larger models when it comes to key criteria and that both model sizes outperform other open models.
In addition to being powerful, Gemma models have been trained to be safe. Google used automated techniques to remove personal information from the data used to train the models, and used reinforcement learning based on human feedback to ensure that variants tailored to Gemma's instructions showed responsible behaviors. Businesses and independent developers could use Gemma to create AI-powered applications, especially if none of the currently available open models are powerful enough for what they want to create.
Google plans to introduce even more variants of Gemma in the future for an even more diverse range of applications. That said, those who want to start working with the models now can access them through the Kaggle data science platform, the company's Colab notebooks, or through Google Cloud.