NTT新成果：All-in-One ASR统一三模型双模式，手机等轻量设备落地更轻松
标题：All-in-One ASR: Unifying Encoder-Decoder Models of CTC, Attention, and Transducer in Dual-Mode ASR 地址：https://arxiv.o…
27 0 1
声浪
![语音/音频处理学术速递[12.19]](https://api.speechhome.com/Resources/Thumb/2025-12-19/2001938819452112896.jpg)
![语音/音频处理学术速递[12.18]](https://api.speechhome.com/Resources/Thumb/2025-12-18/2001586773578354688.jpg)
![语音/音频处理学术速递[12.17]](https://api.speechhome.com/Resources/Thumb/2025-12-17/2001224300123787264.jpg)
![语音/音频处理学术速递[12.16]](https://api.speechhome.com/Resources/Thumb/2025-12-16/2000855943247695872.jpg)
通义开源 0.5B TTS + 0.8B ASR，支持跨语种音色克隆、说唱识别！
只需3秒录音，就能让你的声音无缝切换语种、方言与情绪——中、粤、日、英、开心、愤怒......9 种通用语言、18种方言，通通搞定！ 一段嘈杂环境下的会议录音，AI 也能毫秒级输出文字，绕口令、RAP、背景音乐干扰，照样精准识别！ 通义百聆…
25 0 0
![语音/音频处理学术速递[12.15]](https://api.speechhome.com/Resources/Thumb/2025-12-16/2000855725173248000.jpg)
![语音/音频处理学术速递[12.12]](https://api.speechhome.com/Resources/Thumb/2025-12-12/1999409560842014720.jpg)
![语音/音频处理学术速递[12.11]](https://api.speechhome.com/Resources/Thumb/2025-12-11/1999048806129340416.jpg)
![语音/音频处理学术速递[12.10]](https://api.speechhome.com/Resources/Thumb/2025-12-10/1998676121746542592.jpg)