TII launches Falcon ASR, a speech model tuned for Arabic and the Emirati dialect
Abu Dhabi's Technology Innovation Institute released Falcon-ASR, a 1.6B-parameter speech recognition model. It reports a 20.92% average word error rate across six Arabic test sets.
The Technology Innovation Institute (TII) in Abu Dhabi introduced Falcon-ASR on October 7, a 1.6-billion-parameter speech recognition model focused on Arabic and particularly the Emirati dialect. It also supports English, French, Spanish and Portuguese.
TII says the model averaged a 20.92% word error rate across six Arabic test sets, against a best published result of 23.17% in the leaderboard snapshot it used, and recorded the lowest error rates among systems it compared on its internal Emirati evaluation. These are TII's own evaluations.
The model was trained on Emirati, Modern Standard Arabic, other Gulf and Arabic dialects, and English, and supports word-level timestamps.