Datasetpublic

pollen-robotics/speech-commands-v0.02

Speech Commands Dataset v0.02 This is a re-hosted copy of the Google Speech Commands v0.02 dataset in Parquet format for compatibility with the Hugging Face Dataset Viewer. ⚠️ Credits This dataset was created by Pete Warden / Google. All credit goes to the original authors and the crowdsourcing contributors. Original source: http://download.tensorflow.org/data/speech_commands_v0.02.tar.gz Paper: Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition… See the full description on the dataset page: https://huggingface.co/datasets/pollen-robotics/speech-commands-v0.02.

sourceHugging Facecc-by-4.0updated 3mo agoView on Hugging Face

0likes1.1kdownloads

Dataset Card

Speech Commands Dataset v0.02

This is a re-hosted copy of the Google Speech Commands v0.02 dataset in Parquet format for compatibility with the Hugging Face Dataset Viewer.

⚠️ Credits

This dataset was created by Pete Warden / Google. All credit goes to the original authors and the crowdsourcing contributors.

—Original source: http://download.tensorflow.org/data/speechcommandsv0.02.tar.gz
—Paper: Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition
—License: Creative Commons BY 4.0

If you use this dataset, please cite:

bibtex

@article{speechcommandsv2,
   author = {{Warden}, P.},
    title = "{Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition}",
  journal = {ArXiv e-prints},
archivePrefix = "arXiv",
   eprint = {1804.03209},
 primaryClass = "cs.CL",
     year = 2018,
    month = apr,
     url = {https://arxiv.org/abs/1804.03209},
}

Dataset Description

A set of one-second .wav audio files (16-bit PCM, 16 kHz, mono), each containing a single spoken English word. 105,829 audio files organized into folders by word label.

Words

20 core command words: yes, no, up, down, left, right, on, off, stop, go, zero, one, two, three, four, five, six, seven, eight, nine

15 auxiliary words: bed, bird, cat, dog, happy, house, marvin, sheila, tree, wow, backward, forward, follow, learn, visual

Background noise: the _background_noise_ folder contains longer audio clips of environmental noise.

Splits

The dataset is split into train/validation/test using the official validation_list.txt and testing_list.txt files included in the original archive. Files not in either list are assigned to the training set. The split is deterministic based on a hash of the speaker ID, ensuring all utterances from a given speaker end up in the same partition.

Columns

Column	Type	Description
`audio`	`struct(bytes, path)`	Audio data as WAV bytes, playable in the Dataset Viewer
`label`	`string`	Word label (folder name)
`file`	`string`	Original file path in the archive
`speaker_id`	`string`	Anonymous speaker identifier (hash from filename)
`utterance_id`	`int32`	Utterance index for this speaker/word combination

Re-hosting rationale

The original google/speech_commands dataset on Hugging Face uses a Python loading script that is no longer supported by the Dataset Viewer. This re-hosted version stores the audio data as Parquet files with native audio columns, making it directly viewable and loadable without custom code.

Generated by ML Intern

This dataset repository was generated by ML Intern, an agent for machine learning research and development on the Hugging Face Hub.

—Try ML Intern: https://smolagents-ml-intern.hf.space
—Source code: https://github.com/huggingface/ml-intern

Usage

python

from datasets import load_dataset

dataset = load_dataset('pollen-robotics/speech-commands-v0.02')