© Unsplash/Mojahid Mottakin
It’s big announcement from Google last December: the firm announced three new Large Language Models (LLM) for Bard. Called Gemini, these come in several variants, depending on the platform and use case. The smallest model intended for running locally on devices like smartphones is called Gemini Nano.
It is found in particular in certain applications of the firm on the Pixel 8 and the Samsung Galaxy S24. At the other end of the spectrum, Gemini Ultra – which will be released later in 2024 – is the most advanced of the firm. It is currently reserved for developers: the latter is the result of a monolithic learning protocol, which allows it to generate text as well as images.
Google launches its Gemini Pro model in France
Gemini Pro is an intermediate model, comparable in precision to the GPT model -4 from ChatGPT. However, it is limited to the generation of text, but Bard can still take over and still generate images with its Imagen 2 model. Google is announcing this Thursday, February 1, 2024 that Bard with Gemini Pro is now available in France and in French.
Google explains that Gemini Pro “is particularly effective for understanding, summarizing, reasoning, coding, and planning”. Google adds that this model respects all of the firm's commitments in terms of responsible development of AI – and emphasizes that “experts and public decision-makers regulating privacy” have had their say on this international expansion.
In addition to this, Google is announcing the wider deployment of a feature called “Check answer”. When Bard responds to a prompt, a button, bearing the letter “G” allows you to search the web in one click to see if there are any results to support this answer. This will make it easier to avoid errors and confusion that sometimes arise from prompts generated by including a web search.
Google also announces the arrival of the model image generation tool Imagen 2 in Bard. What to combine according to the firm “quality and speed, delivering high-quality photo-realistic results”
How to test Bard with Gemini Pro?
Bard is available on the web. To test the new model, simply go to https://bard.google.com/. Gemini Pro is currently being deployed, so the new model may not be available to you immediately.
For those who are more impatient, it is possible to test Gemini Pro immediately by changing the language of your Google account:
- Go to the page Manage your account Google.
- Select Data and customization.
- Under Language and region, click “Language.
- Select English (US).
- Open the Bard website: https://bard.google.com/
- Log in to your Google account.
You should see a blue banner indicating that Bard has been updated with Gemini Pro. Note that even if you have Gemini Pro on your Google account in French, you will only be able to generate images with Imagen 2 with prompts in English.
📍 To not miss any news from Presse-citron, follow us on Google News and WhatsApp.
156.9 M reviews