Millions of spoken keywords for developing keyword spotters!
Get StartedKeyword spotting (KWS) has become a hot topic in speech processing due to the rise of commercial applications based on voice command detection, such as voice assistants. This project demonstrates how anyone can reproduce SiDi KWS, a large-scale multilingual dataset of spoken keywords, by running open-source automatic forced alignment tools with public datasets.
We are at INTERSPEECH 2022!More than 24.3 million labeled audio clips of spoken keywords!
Around 700 thousand unique keywords.
Keywords in English, French, German and Spanish!
Based on public transcribed speech datasets and aligner.
Use any programming language to segment the input speech.
Enjoy your new large-scale keyword spotting dataset!
SiDi KWS has been developed at SiDi as an internal research project conducted by the following authors:
SiDi KWS has been financed by Samsung Eletrônica da Amazonia Ltda., under the auspices of the Brazilian Federal Law of Informatics nº. 8248/91. Due to a coorporate policy, neither the original SiDi KWS dataset nor the source code used to build it can be made public.
