ai4bharat/Rural_Women_ASR_v2

Rural Women ASR Dataset (Hindi & Bhojpuri)

The Rural Women ASR Dataset is part of the Recognizing Every Voice initiative, which focuses on building inclusive Automatic Speech Recognition (ASR) systems for rural women in India.

It includes Hindi and Bhojpuri speech data collected from rural women speakers, covering diverse age groups, regions, and socio-economic backgrounds.

---

🧩 Dataset Summary

Feature	Description
language	Language of the utterance (Hindi or Bhojpuri)
audio_path	.wav audio file
text	Transcribed text for the audio
verbatim	Raw, unedited transcription as spoken
normalized	Cleaned and normalized transcription
audio_raw_duration	Duration of original audio file in seconds
audio_chunk_duration	Duration of segmented audio chunk in seconds
scenario	Read, Extempore
task_name	Task label for data collection
speaker_id	Unique speaker identifier
gender	Gender of the speaker (majority: Female)
age_group	Age group of the participant
occupation	Occupation of the speaker
qualification	Education level
area	Rural/urban classification
district	District of the speaker
state	State of the speaker
verification_report	Quality/validation notes
prompt	Prompt or seed text used while recording (if applicable)

---

🌍 Dataset Details

• Languages: Hindi, Bhojpuri
• Regions Covered: Rural districts in Madhya Pradesh, Uttar Pradesh and Bihar.
• Speakers: Rural women across diverse job types (farmers, homemakers, artisans, etc.)
• Collection Method: Field recordings using mobile phones via KathBath application
• Sampling Rate: 48kHz WAV files
• Transcriptions: Human-verified text transcriptions

---

---

Splits

Split	Language	Examples	Size (approx.)
Hindi	Hindi	30,827	31.38 GB
Bhojpuri	Bhojpuri	33,616	34.22 GB

---

How to Use

The dataset can be easily loaded using the Hugging Face datasets library.

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("ai4bharat/Rural_Women_ASR_v2")

# Access a specific split (e.g., Hindi or Bhojpuri)
hindi_split = dataset["Hindi"]
bhojpuri_split = dataset["Bhojpuri"]

# Print the first example
print(bhojpuri_split[0])

Citation

If you use this dataset in your research, please cite the following paper:

@misc{joshi2025recognizingvoiceinclusiveasr,
      title={Recognizing Every Voice: Towards Inclusive ASR for Rural Bhojpuri Women}, 
      author={Sakshi Joshi and Eldho Ittan George and Tahir Javed and Kaushal Bhogale and Nikhil Narasimhan and Mitesh M. Khapra},
      year={2025},
      eprint={2506.09653},
      archivePrefix={arXiv},
      primaryClass={eess.AS},
      url={[https://arxiv.org/abs/2506.09653](https://arxiv.org/abs/2506.09653)}, 
}

Licensing

Released under the Creative Commons Attribution 4.0 (CC-BY-4.0) License.
You are free to use, share, and adapt with attribution.

---