Post
702
Xiaohongshu (RedNote) has quietly shared an open TTS - FireRedTTS 2 on the hub 🔥
FireRedTeam/FireRedTTS2
FireRedTeam/FireRedTTS2
✨ Long conversational TTS → 3 min, 4 speakers, scalable to more
✨ Multilingual & cross-lingual voice cloning (English, Chinese, Japanese, Korean, French, German, Russian)
✨ Ultra-low latency → 140ms first-packet on L20 GPU
✨ Random timbre generation for ASR/interaction data
FireRedTeam/FireRedTTS2
FireRedTeam/FireRedTTS2
✨ Long conversational TTS → 3 min, 4 speakers, scalable to more
✨ Multilingual & cross-lingual voice cloning (English, Chinese, Japanese, Korean, French, German, Russian)
✨ Ultra-low latency → 140ms first-packet on L20 GPU
✨ Random timbre generation for ASR/interaction data