zai-org/GLM-ASR
807
GLM-ASR-Nano: A robust, open-source speech recognition model with 1.5B parameters
鲁棒的开源语音识别模型，含15亿参数
FunAudioLLM/Fun-ASR
1,202
End-to-end speech recognition large model: 31 languages, dialects, accents, lyrics, hotwords, timestamps, speaker diarization. Trained on tens of millions of hours.
支持31种语言、方言、歌词、热词、时间戳和说话人日志的端到端语音识别大模型。
FunAudioLLM/SenseVoice
8,411
Multilingual speech understanding: ASR + emotion recognition + audio event detection. 50+ languages, 15x faster than Whisper, non-autoregressive.
多语言语音理解，支持ASR、情感识别和音频事件检测，速度比Whisper快15倍