🎙️ SenseVoice

Speech Recognition + Emotion Detection + Audio Events — All in One Model

5 languages (zh/en/yue/ja/ko) · 7x faster than Whisper-small · 17x faster than Whisper-large

⭐ GitHub · 🛠️ FunASR Toolkit · 📄 Paper · 🚀 Fun-ASR (31 Languages)

How it works: Upload audio or record via microphone → SenseVoice transcribes speech and detects emotions (😊😡😔) and sound events (🎼👏😀😭🤧). Event labels appear at the front of text, emotions at the end.