README.md

metadata

license: mit
language:
  - rw
pretty_name: KinyaWhisper
task_categories:
  - automatic-speech-recognition
tags:
  - rw
  - Kin
  - Kinyarwanda
  - Rwanda
  - ururimi
  - rwacu
  - ururimi rwacu
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: text
      dtype: string
    - name: audio_len
      dtype: float64
    - name: transcript_len
      dtype: int64
    - name: len_ratio
      dtype: float64
  splits:
    - name: train
      num_bytes: 2475947
      num_examples: 102
  download_size: 2478359
  dataset_size: 2475947

Kinyarwanda Spoken Words Dataset

This dataset contains 102 short audio samples of spoken Kinyarwanda words, each labeled with its corresponding transcription. It is designed for training, evaluating, and experimenting with Automatic Speech Recognition (ASR) models in low-resource settings.

Structure

• audio/: Contains 102 .wav files (mono, 16kHz)
• transcripts.txt: Tab-separated transcription file (e.g., 001.wav\tmuraho)
• manifest.jsonl: JSONL file with audio paths and text labels (compatible with 🤗 Datasets and Whisper training scripts)

Example

{"audio_filepath": "audio/001.wav", "text": "muraho"}

Usage

from datasets import load_dataset

ds = load_dataset("benax-rw/my_kinyarwanda_dataset", split="train")
example = ds[0]
print(example["audio"]["array"], example["text"])

License

This dataset is published for educational and research purposes.

Citation

If you use this dataset, please cite:

Benax Labs, KinyaWhisper Dataset for Fine-tuning Whisper on Kinyarwanda (2025)