Bhashini data to be announced in Oct: CEO
NEW DELHI : Bhashini, the government’s national language translation project, will release data on early adoption among private companies, as well as on the size of repositories on Indian languages collected and generated by the project in October this year, said Amitabh Nag, chief executive of Bhashini. “The internet does not have data repositories in languages such as Manipuri, which Bhashini is now building. It is still early days for the size of repositories in certain languages, but Bhashini will be making an announcement on the details of this by October this year," Nag told Mint . The executive affirmed that Bhashini now has data for text-to-text translations in 22 languages, while data for voice-to-text translations are available in 14 languages.
With increasing repositories, adoption is also growing. “Private entities have already expressed interest in adopting Bhashini. You will see these announcements coming in gradually by October, and we are presently at a proof-of-concept stage for private entities," Nag added.
Bhashini aims to create a local language repository on the internet, which can be accessed by various stakeholders. On 28 June. Manish Gupta, director of Google Research India, told Mint that Google is building Project Vaani to make local Indian language data available to researchers and academics across the country.
Industrial adoption of this will follow. On 9 September, Mint reported Reliance Industries’ Jio Platforms’ partnership with Nvidia to build an AI infrastructure that accelerates development and building of local large language models for industrial generative artificial intelligence use. Such adoptions are gradually increasing with the Centre’s repository as well, going beyond
. Read more on livemint.com
