mteb/SpeechCommandsZeroshotv0.02

SpeechCommandsZeroshotv0.02 An MTEB dataset Massive Text Embedding Benchmark Sound Classification/Keyword Spotting Dataset. This is a set of one-second audio clips containing a single spoken English word or background noise. These words are from a small set of commands such as 'yes', 'no', and 'stop' spoken by various speakers. With a total of 10 labels/commands for keyword spotting and a total of 30 labels for other auxiliary tasks Task category a2t Domains Spoken… See the full description on the dataset page: https://huggingface.co/datasets/mteb/SpeechCommandsZeroshotv0.02.