Home BlogDataset Teaching Machines How People Talk

Teaching Machines How People Talk

by Michael McLaughlin
Sound waves.

Mozilla has published the latest dataset from its Common Voice project, which aims to spur the development of voice-enabled technologies. The dataset consists of nearly 1,400 hours of recordings from 42,000 individuals speaking a total of 18 different languages. In addition, the dataset includes labels such as the age, sex, and accent of contributors who opted in to provide the metadata.

Get the data.

Image: DPic

You may also like

Show Buttons
Hide Buttons