Google is Google, a year after the release of two “Open” JEMA AI models made from the same technology behind the Gemini AI, is Google Updating the family with JEMA 3. According to the blog post, the purpose of these models is to use by developers who manufacture AI applications wherever needed, with the ability to run from phone to more than 35 languages, as well as the ability to analyze text, images and short videos.
The company claims that it is a “world’s best only exclusive model” from Facebook’s Lama, Deep Sak, and Open AI for performance with a GPU, as well as better capabilities to run on NVIDIA’s GPU and dedicated AI hardware as well as improve better abilities. The vision encoder of JEMA 3 has also been upgraded with high -river and non -square images support, while the new Shieldjama 2 Image Safety Rating Available to filter both image input and output for sexually clear, dangerous, or violent content.
Last year, it was unclear how much interest would be in a model like Jema, however, the popularity of Deep Sak and others showed that AI -tech is interested in hardware with low requirements. Despite the claims of advanced abilities, Google also says, “Gema 3’s improved stem performance gave rise to specific diagnosis focusing on the possibility of misuse in producing harmful substances. Their results indicate a lower level of risk.
Exactly what is the “open” or “open source” AI model, and with Google’s Jema, it focuses on the company’s license that allows people to use it for it, which has not changed with this new release. Google continues to promote JEMA with Google Cloud Credit, and the Jima 3 educational program will allow academic researchers to apply for $ 10,000 worth of credit to accelerate their research.
