Google Releases Gemma Family of Open Source AI Models for AI Developers and Researchers. The company told about this in its blog.
They are built on the same technology and research as Gemini.
Gemma comes in two sizes – the lightweight 2B and the more productive 7B. Google claims that both variants achieve the best performance compared to other open models in their class.
Colab and Kaggle services are available for users to use, as well as integration with popular tools such as Hugging Face, MaxText, NVIDIA NeMo and TensorRT-LLM.
Pre-trained and pre-configured Gemma models can run on a laptop or in the cloud. The company emphasized that the development was optimized for NVIDIA graphics processors.
Gemma models can be used in commercial projects, regardless of their size.
Gemma was developed by DeepMind and other Google teams, and its name reflects the Latin word gemma, which means precious stone.