CASPER is a speech dataset comprised of recordings of two people engaging in spontaneous conversations in English. The dataset aims to fill the gap in high quality spontaneous speech data. The conversations were conducted over a custom-built web platform from each participant's end and their own device, fostering genuine interactions and natural conversations with a diverse range of topics.
How to Download the Dataset… See the full description on the dataset page: https://huggingface.co/datasets/CASPER-SSSD/CASPER.
