Hugging Face Blog·· 2024-10-22AI score40
How Hugging Face deploys Speech-to-Speech on Inference Endpoints using custom Docker images
Deploying Speech-to-Speech on Hugging Face
AI briefing
Hugging Face's Speech-to-Speech (S2S) project implements voice conversation through a cascaded pipeline consisting of VAD, STT, language model, and TTS, supporting English, French, Spanish, Chinese, Japanese, and Korean. It can run in a single language or use the auto flag to automatically detect the language.
Open original huggingface.co