Confirmed
Tencent releases Hy-SimulMT, a 30B real-time speech translation model for Chinese and English
Hy-SimulMT translates speech to speech between Chinese and English as the speaker talks, with real-time voice cloning. Its RealSI evaluation shows about 3 seconds StreamLAAL, and GGUF versions are on Hugging Face.
Every outlet that ran it 1 article
r/LocalLLaMACommunityLead