Google releases Gemma 2 language model with improved performance, inference efficiency, and security

Google announced the release of its Gemma 2 language model for researchers and developers around the world. According to reports, Gemma 2 has two parameter sizes of 9 billion (9B) and 27 billion (27B). Compared with the first generation, it has higher performance, more efficient inference, and has significant built-in security improvements. Google says the Gemma 2 27B's performance is more competitive than similar products twice its size; the 9B's performance is also class-leading, outperforming the Llama 3 8B and other open models. (Interface News)