Google releases Gemma 7B, 2B open-source AI models

MANILA, Philippines – Google on Wednesday, February 21, announced it had released two new open-source artificial intelligence models built from “the same research and technology used to create the Gemini models.”

The two models, Gemma 7B and Gemma 2B, refer to the “7 billion parameter model for efficient deployment and development on GPU and TPU, and a 2 billion parameter model for CPU and on-device applications,” Google said in its technical report on Gemma.

The models were released with pre-trained and instruction-tuned versions, and Google said these variants can “run on your laptop, workstation, or Google Cloud with easy deployment on Vertex AI and Google Kubernetes Engine (GKE).”

Gemma 7B and 2B are available worldwide via Kaggle, Hugging Face, Nvidia NeMo, and Vertex AI.

The Gemma models offer text-to-text support only, unlike the Gemini model, but are said to surpass “significantly larger models on key benchmarks while adhering to our rigorous standards for safe and responsible output.”

Alongside the Gemma models, Google is also releasing a responsible generative AI toolkit to support best practices in using and working on the AI models.

The news follows Google bringing Gemini AI to its enterprise workspace tools. – Rappler.com