Tag
NVIDIA's entire speech stack—ASR, TTS, and codec—is now quantized to GGUF and runs locally on-device via NeMo-Speech.cpp, with new model releases for Magpie-TTS, Nemotron Speech Streaming, and Parakeet.