A new FFASR leaderboard ranks automatic speech recognition models on messy, real-world audio instead of clean lab test sets.
Continue to AI University →