Dataset or benchmark
speech_commands dataset card
Hugging Face, dataset card for Google's Speech Commands dataset, 2018
Read the originalRead on October 17, 2026
- What it is
- The dataset card for Google's Speech Commands, a set of one-second recordings of single spoken words and background noise, built for training simple keyword-spotting models. It lists each version's release date and file count.
- How much weight it carries
- Maintained with the dataset and linked to the arXiv paper that describes the collection. It gives file counts, not speaker counts or hours, and it covers single words, not sentences, so it is a keyword benchmark rather than a dictation corpus.
Facts taken from this source
- Version 0.02 of Google's Speech Commands dataset contains 105,829 audio files, each a one-second recording of a single spoken English word or background noise.
“each containing a single spoken English word or background noise. These words are from a small set of commands ... Version 0.02 of the data set (configuration "v0.02" ) was released on April 11th 2018 and contains 105,829 audio files.” (Dataset description and Dataset Summary)