Hugging Face's logo

Datasets:
ai4bharat
/
Shrutilipi

Modalities:
Audio
Text
Formats:
parquet
Size:
1M - 10M
ArXiv:
Libraries:
Datasets
Dask
License:
Dataset card Data Studio Files Community
2
Shrutilipi / README.md
tahirjm's picture
tahirjm
Update README.md
f4beb2f verified
preview code
|
raw
history blame contribute delete
9.35 kB
metadata
license: cc-by-4.0
dataset_info:
  - config_name: assamese
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 5082850536.85
        num_examples: 41766
    download_size: 5229826172
    dataset_size: 5082850536.85
  - config_name: bengali
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 1970326408.576
        num_examples: 17882
    download_size: 1916156215
    dataset_size: 1970326408.576
  - config_name: dogri
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 3603495138.825
        num_examples: 23721
    download_size: 3568871899
    dataset_size: 3603495138.825
  - config_name: gujarati
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 16321596906.35
        num_examples: 137965
    download_size: 17665902401
    dataset_size: 16321596906.35
  - config_name: hindi
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 97821772346.85
        num_examples: 734675
    download_size: 94834500694
    dataset_size: 97821772346.85
  - config_name: kannada
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 20069410879.104
        num_examples: 192208
    download_size: 21641876571
    dataset_size: 20069410879.104
  - config_name: konkani
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 1246784105.294
        num_examples: 12239
    download_size: 1239977556
    dataset_size: 1246784105.294
  - config_name: maithili
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 2985296990.293
        num_examples: 22233
    download_size: 2895972814
    dataset_size: 2985296990.293
  - config_name: malayalam
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 17608171088.553
        num_examples: 224323
    download_size: 19431260422
    dataset_size: 17608171088.553
  - config_name: marathi
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 42353055748.775
        num_examples: 336475
    download_size: 38603401934
    dataset_size: 42353055748.775
  - config_name: nepali
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 5214995797.048
        num_examples: 31997
    download_size: 4990235585
    dataset_size: 5214995797.048
  - config_name: odia
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 12687640571.75
        num_examples: 126675
    download_size: 16473690962
    dataset_size: 12687640571.75
  - config_name: punjabi
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 3054398341.146
        num_examples: 21118
    download_size: 3181612927
    dataset_size: 3054398341.146
  - config_name: sanskrit
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 2036760741.337
        num_examples: 14177
    download_size: 2036474974
    dataset_size: 2036760741.337
  - config_name: tamil
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 33855148157.352
        num_examples: 281508
    download_size: 31654328747
    dataset_size: 33855148157.352
  - config_name: telugu
    features:
      - name: audio_filepath
        dtype: audio
      - name: text
        dtype: string
      - name: duration
        dtype: float64
      - name: lang
        dtype: string
    splits:
      - name: train
        num_bytes: 5722466966.018
        num_examples: 56082
    download_size: 5886818756
    dataset_size: 5722466966.018
configs:
  - config_name: assamese
    data_files:
      - split: train
        path: assamese/train-*
  - config_name: bengali
    data_files:
      - split: train
        path: bengali/train-*
  - config_name: dogri
    data_files:
      - split: train
        path: dogri/train-*
  - config_name: gujarati
    data_files:
      - split: train
        path: gujarati/train-*
  - config_name: hindi
    data_files:
      - split: train
        path: hindi/train-*
  - config_name: kannada
    data_files:
      - split: train
        path: kannada/train-*
  - config_name: konkani
    data_files:
      - split: train
        path: konkani/train-*
  - config_name: maithili
    data_files:
      - split: train
        path: maithili/train-*
  - config_name: malayalam
    data_files:
      - split: train
        path: malayalam/train-*
  - config_name: marathi
    data_files:
      - split: train
        path: marathi/train-*
  - config_name: nepali
    data_files:
      - split: train
        path: nepali/train-*
  - config_name: odia
    data_files:
      - split: train
        path: odia/train-*
  - config_name: punjabi
    data_files:
      - split: train
        path: punjabi/train-*
  - config_name: sanskrit
    data_files:
      - split: train
        path: sanskrit/train-*
  - config_name: tamil
    data_files:
      - split: train
        path: tamil/train-*
  - config_name: telugu
    data_files:
      - split: train
        path: telugu/train-*

Shrutilipi

GitHub ArXiv CC BY 4.0

Dataset Description

Overview

Shrutilipi is a labelled ASR corpus obtained by mining parallel audio and text pairs at the document scale from All India Radio news bulletins for 12 Indian languages: Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Odia, Punjabi, Sanskrit, Tamil, Telugu, Urdu. The corpus has over 6400 hours of data across all languages.

This work is funded by Bhashini, MeitY and Nilekani Philanthropies

Usage

The datasets library enables you to load and preprocess the dataset directly in Python. Ensure you have an active HuggingFace access token (obtainable from Hugging Face settings) before proceeding.

To load the dataset, run:

from datasets import load_dataset
# Load the dataset from the HuggingFace Hub
dataset = load_dataset("ai4bharat/Shrutilipi","bengali",split="train")
# Check the dataset structure
print(dataset)

You can also stream the dataset by enabling the streaming=True flag:

from datasets import load_dataset
dataset = load_dataset("ai4bharat/Shrutilipi","bengali",split="train", streaming=True)
print(next(iter(dataset)))

Citation

If you use Shrutilipi in your work, please cite us:

@inproceedings{DBLP:conf/icassp/BhogaleRJDKKK23,
  author       = {Kaushal Santosh Bhogale and
                  Abhigyan Raman and
                  Tahir Javed and
                  Sumanth Doddapaneni and
                  Anoop Kunchukuttan and
                  Pratyush Kumar and
                  Mitesh M. Khapra},
  title        = {Effectiveness of Mining Audio and Text Pairs from Public Data for
                  Improving {ASR} Systems for Low-Resource Languages},
  booktitle    = {{ICASSP}},
  pages        = {1--5},
  publisher    = {{IEEE}},
  year         = {2023}
}

License

This dataset is released under the CC BY 4.0.

Contact

For any questions or feedback, please contact: