Gemma 3 270M: The compact model for hyper-efficient AI
https://developers.googleblog.com/en/introducing-gemma-3-270m/