Hi all. Except the single word dataset from common voice, and Pete’s speech command dataset, is there any other dataset suitable for similar purpose? Since the keywords I desired are not in these 2 datasets.
The keywords I want are “flush” and “help”, is finding someone to record audio the only way I can achieve this?
The new work out of VJ’s lab is probably the biggest example of a dataset: GitHub - harvard-edge/multilingual_kws: Few-shot Keyword Spotting in Any Language and Multilingual Spoken Word Corpus
1 Like