A Fundamental End-to-End Speech Recognition Toolkit and Open Source SOTA Pretrained Models, Supporting Speech Recognition, Voice Activity Detection, Text Post-processing etc.

SpeechGPT-2.0-preview is an advanced natural language processing model specifically designed for generating coherent and realistic speech patterns. This cutting-edge tool leverages state-of-the-art deep learning techniques to enable more accurate and human-like speech synthesis.

FunASR实时语音识别版，识别麦克风和电脑内播放的声音，电脑语音打字软件