This is the repository for the following papers:
* [Few-Shot Keyword Spotting in Any Language](https://www.isca-archive.org/interspeech_2021/mazumder21_interspeech.pdf), Interspeech 2021
* [Multilingual Spoken Words Corpus](https://openreview.net/pdf?id=c20jiJ5K2H), NeurIPS 2021 Datasets and Benchmarks Track


# Introductory Colab Tutorial

Launch [our introductory tutorial on Google Colab](https://colab.research.google.com/github/harvard-edge/multilingual_kws/blob/main/multilingual_kws_intro_tutorial.ipynb) to explore our dataset and train a few-shot keyword-spotting model:

<a href="https://colab.research.google.com/github/harvard-edge/multilingual_kws/blob/main/multilingual_kws_intro_tutorial.ipynb"><img src="https://www.tensorflow.org/images/colab_logo_32px.png" alt="https://colab.research.google.com/github/harvard-edge/multilingual_kws/blob/main/multilingual_kws_intro_tutorial.ipynb"></a>

# Download the Multilingual Spoken Words Corpus

Our full dataset is available at https://mlcommons.org/words

# Videos

[[`Few-Shot Keyword Spotting in Any Language: youtube video overview`](https://www.youtube.com/watch?v=1JRlJoEhYxE)]

[![](images/interspeech21_multilingual_kws.png)](https://www.youtube.com/watch?v=1JRlJoEhYxE)

[[`Multilingual Spoken Words Corpus: youtube video overview`](https://www.youtube.com/watch?v=eGPCwnRtRng)]

[![](images/neurips21_mswc.png)](https://www.youtube.com/watch?v=eGPCwnRtRng)

---