Google recently released Gemma, a new open source large language model series, whose performance surpasses existing models such as Llama 2 and Mistral. Gemma uses a unique GeGLU activation function and has made breakthroughs in parameter scale, showing strong potential to replace existing open source models. This release marks another major development in the field of open source AI, providing new options for developers and researchers.
Google has launched a new open source model series "Gemma". Gemma performs well compared with Llama 2 and Mistral and has unique design principles. Gemma uses the GeGLU activation function, which has a large parameter scale but can replace other open source models, showing strong potential.
The emergence of Gemma has injected new vitality into the open source AI field, and its performance advantages and unique design deserve attention. In the future, we look forward to Gemma showing its potential in more application scenarios and promoting the development of open source AI technology.