Rohan Paul· @rohanpaul_ai · X·· 2 小时前AI 评分52
AI 导读
@shobhitbanga 发布 Joy，一个双通道对话数据集，包含 10 万小时、5 万说话人、10 种语言的真实非脚本双人对话，用于训练语音模型理解何时倾听、何时插话。Rohan Paul 转发并指出当前语音 AI 多从 LLM 生成、TTS 朗读的假对话学习，听不到人们互相抢话，同时好奇对话如何维持 10 到 20 分钟。配图显示 Joy 平均每小时说话人重叠约 350 次、backchannel 约 300 次。
正文 · 原文
Unscripted calls are hard to collect at scale.
Voice AI today mostly learns from fake dialogue written by LLMs and read aloud by TTS, so it never hears people talking over each other;
Really curious how they kept the conversations going for 10 to 20 minutes.
They released Joy, a 100,000 hours of unscripted, two-person conversation across 10 languages
https://x.com/shobhitbanga/status/2115862857351061694/video/1
Drop 1/15: Introducing Joy, a dual-channel conversational dataset. Real, unscripted conversations between two people, each on their own track. Built to teach voice models how people actually talk: when to listen, when to jump in, when to say “hmm”. 100,000 hours. 50,000 speakers. 10 languages.在 X 查看被引用的帖子
来源：Rohan Paul · x.com