Digital Divide Data

license

cc-by-sa-4.0

access

public(ungated on Hugging Face

dataset_card

present(speakers, topics, format and sample rows)

The audio and transcripts download without a gate under a share-alike license that asks users to credit Digital Divide Data. The card describes speakers, topics and file layout, though no paper accompanies it.

• https://huggingface.co/api/datasets/Digital-Divide-Data/khmer-speech-dataset recorded 2026-09-24

  gated: false; license:cc-by-sa-4.0

• https://huggingface.co/datasets/Digital-Divide-Data/khmer-speech-dataset/raw/main/README.md recorded 2026-09-24

  "license is Creative Commons Attribution Share Alike 4.0 International (CC-BY-SA-4.0). Please attribute Digital Divide Data"

Adoption is Hugging Face downloads of the repository. Copies taken from the Mozilla Data Collective listing are not included.

• https://huggingface.co/api/datasets/Digital-Divide-Data/khmer-speech-dataset recorded 2026-09-24

  3787 downloads in the trailing 30 days for Digital-Divide-Data/khmer-speech-dataset

The corpus dwarfs OpenSLR's earlier Khmer speech-text set of under four hours, pairing prompted readings with curated transcripts, and community Khmer recognizers built on Whisper and Qwen are trained on it. At about seven hundred hours from twelve speakers reading prepared sentences it is small beside English speech corpora of hundreds of thousands of hours, and IndicVoices draws spontaneous speech from tens of thousands of speakers.

• https://huggingface.co/datasets/Digital-Divide-Data/khmer-speech-dataset recorded 2026-09-24

  Number of rows: 450,396; Models trained or fine-tuned: seanghay/Qwen3-ASR-0.6B-Khmer, sengtha/whisper-base-khmer

• https://huggingface.co/datasets/Digital-Divide-Data/khmer-speech-dataset/raw/main/README.md recorded 2026-09-24

  "727.94 hours of manually curated speech-text pairs by native speakers"; "there was only one Khmer speech-text dataset: OpenSLR 42 with 3.97 hours"