Skip to content

Dataset · artur-muratov

Multilingual speech commands (15 languages, augmented)

artur-muratov/multilingual-speech-commands-15lang

Augmented audio recordings of short voice commands in 15 languages for keyword spotting systems.

What's inside

The dataset includes 1M to 10M audio-text pairs in 15 languages, including English, Russian, Kazakh, Tatar, Arabic, and Turkish. Recordings are collected from various open sources and filtered to match the Google Speech Commands vocabulary. Standard augmentation methods are applied to improve model robustness to audio recordings.

What you can do with it

  • Training compact Keyword Spotting models for smart devices
  • Fine-tuning multilingual ASR models on Turkic and low-resource languages (Kazakh, Tatar)
Languages
English, Russian, Kazakh, Tatar, Arabic, Turkish, French, German and 4 more
Size
1M–10M rows
Downloads this month
20.6K
Likes
16

How to load it

from datasets import load_dataset

ds = load_dataset("artur-muratov/multilingual-speech-commands-15lang")