Meta 推出 Muse Voice Transcribe,支持说话人分离,每小时 0.18 美元

该模型可处理20多位说话者并支持语码切换,而 Speechmatics 和 AWS 则公布了更高的说话者上限。

By · Published

Professionals in a modern meeting room, with their spoken words visualized as color-coded text streams organizing on a translucent table.

该模型可处理20多位说话者并支持语码切换,而 Speechmatics 和 AWS 则公布了更高的说话者上限。