Overview

• The model, called SeamlessM4T, can translate between speech and text for up to 100 input languages and 35 output languages.
• SeamlessM4T is the first single AI model capable of speech-to-text, text-to-speech, speech-to-speech, and text-to-text translation.
• Meta aims for SeamlessM4T to enable seamless communication across languages, likening it to the universal Babel fish translator in science fiction.
• SeamlessM4T outperforms previous models in low and mid-resource languages while maintaining accuracy in high-resource languages.
• Meta is releasing SeamlessM4T and its training dataset SeamlessAlign publicly to advance multilingual AI research.