Jev Collection

Jev Collection is a dataset of 25,000 hours of labeled audio recordings of individuals speaking in 11 languages, collected from the internet and crowdsourced. The dataset is designed to be used for speech recognition and other natural language processing tasks. The data was collected by a company called Jev, which is a data annotation platform. The dataset includes a wide range of speakers, ages, and accents, making it suitable for training AI models that need to handle diverse speech patterns. The dataset is available for free, but users need to create an account on the Jev platform to access it.