otoSpeech Task Dataset Released by otoDotEarth

The account @otodotearth posted that it is releasing otoSpeech Task, described as a task-oriented conversation dataset consisting of 20 hours of English two-person task conversation. Researcher Kyle Kastner at Google DeepMind, who works on machine learning for audio, speech synthesis, and music generation, retweeted the post. The announcement appears in the public posts shared in the visible conversation. No further details on access, licensing, or specific task types are stated in the source lines.