News Post

Meta’s “massively multilingual” AI model translates up to 100 languages, speech or text

ArstechnicaArstechnica · 1y
Meta’s “massively multilingual” AI model translates up to 100 languages, speech or text

Meta has unveiled SeamlessM4T, a multimodal AI model capable of text-to-speech, speech-to-text, speech-to-speech, and text-to-text translations for approximately 100 languages. Meta is releasing SeamlessM4T under a research license, allowing developers to build on the work. It is also offering SeamlessAlign, described as "the biggest open multimodal translation dataset to date," containing 270,000 hours of speech and text alignments. This release is part of Meta's effort to improve language translation and make communication easier across various languages and modalities, aligning with its vision of a universal language translator akin to the Babel Fish from "The Hitchhiker's Guide to the Galaxy."

Comments

Download the medial app to read full posts, comements and news.