From the source
Hugging Face and TII introduced Falcon-ASR, a 1.6 billion parameter speech recognition model for Arabic with a focus on the Emirati dialect, also supporting English, French, Spanish, and Portuguese.
In evaluations, it achieved an average word error rate of 20.92% across six Arabic test sets, outperforming the best published result of 23.17% in the leaderboard snapshot used.
The model supports word-level timestamps and is available via a Hugging Face demo, with API access and native applications planned.