espressif/esp-sr
Speech recognition
ESP-SR ─ Speech framework for ESP32 chips
├── Wake Word Engine (WakeNet) ─ always-on keyword detection
│ └─ models: wn9, wn9s, wn10 … "Hi ESP", "Alexa", multilingual
├── Speech Command Recognition (MultiNet) ─ up to 300 offline commands
│ └─ Chinese + English, no retraining to add commands
├── Audio Front-End (AFE) ─ AEC, VAD, BSS, NS, deep noise suppression
├── VADNet ─ voice-activity detection model (replaces WebRTC VAD)
└── Speech Synthesis (TTS) ─ Pipeline V3, multilingual wake-word training
관련
- 프로젝트
- 프로젝트
- 프로젝트
- 프로젝트
- 프로젝트