README.md

metadata

language:
  - as
  - awa
  - bho
  - brx
  - doi
  - grt
  - ks
  - kok
  - kru
  - mai
  - mni
  - mwr
  - ne
  - or
  - sat
  - sd
license: cc-by-4.0
task_categories:
  - automatic-speech-recognition
pretty_name: Indic Dialect ASR
size_categories:
  - 1M<n<10M
dataset_info:
  - config_name: bhojpuri
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: sentence
        dtype: string
      - name: language
        dtype: string
      - name: source
        dtype: string
    splits:
      - name: train
        num_bytes: 27960185475.376
        num_examples: 95672
    download_size: 16621044137
    dataset_size: 27960185475.376
  - config_name: chhattisgarhi
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: sentence
        dtype: string
      - name: language
        dtype: string
      - name: source
        dtype: string
    splits:
      - name: train
        num_bytes: 23497108435.004
        num_examples: 85793
    download_size: 19890873718
    dataset_size: 23497108435.004
  - config_name: harouti
    features:
      - name: audio
        dtype: audio
      - name: sentence
        dtype: string
      - name: language
        dtype: string
      - name: source
        dtype: string
    splits:
      - name: train
        num_examples: 11
  - config_name: magahi
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: sentence
        dtype: string
      - name: language
        dtype: string
      - name: source
        dtype: string
    splits:
      - name: train
        num_bytes: 19938507749
        num_examples: 95275
    download_size: 17916848914
    dataset_size: 19938507749
  - config_name: rajasthani
    features:
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: sentence
        dtype: string
      - name: language
        dtype: string
      - name: source
        dtype: string
    splits:
      - name: train
        num_bytes: 2467499534.712
        num_examples: 12208
    download_size: 2832188901
    dataset_size: 2467499534.712
default_config_name: harouti
configs:
  - config_name: assamese
    data_files:
      - split: train
        path: assamese/train-*.parquet
  - config_name: awadhi
    data_files:
      - split: train
        path: awadhi/train-*.parquet
  - config_name: bagri
    data_files:
      - split: train
        path: bagri/train-*.parquet
  - config_name: bhili
    data_files:
      - split: train
        path: bhili/train-*.parquet
  - config_name: bhojpuri
    data_files:
      - split: train
        path: bhojpuri/train-*
  - config_name: bodo
    data_files:
      - split: train
        path: bodo/train-*.parquet
  - config_name: braj
    data_files:
      - split: train
        path: braj/train-*.parquet
  - config_name: bundeli
    data_files:
      - split: train
        path: bundeli/train-*.parquet
  - config_name: chhattisgarhi
    data_files:
      - split: train
        path: chhattisgarhi/train-*
  - config_name: dogri
    data_files:
      - split: train
        path: dogri/train-*.parquet
  - config_name: garhwali
    data_files:
      - split: train
        path: garhwali/train-*.parquet
  - config_name: garo
    data_files:
      - split: train
        path: garo/train-*.parquet
  - config_name: harouti
    data_files:
      - split: train
        path: harouti/train-*.parquet
  - config_name: haryanvi
    data_files:
      - split: train
        path: haryanvi/train-*.parquet
  - config_name: kangri
    data_files:
      - split: train
        path: kangri/train-*.parquet
  - config_name: kashmiri
    data_files:
      - split: train
        path: kashmiri/train-*.parquet
  - config_name: konkani
    data_files:
      - split: train
        path: konkani/train-*.parquet
  - config_name: kumaoni
    data_files:
      - split: train
        path: kumaoni/train-*.parquet
  - config_name: kurukh
    data_files:
      - split: train
        path: kurukh/train-*.parquet
  - config_name: magahi
    data_files:
      - split: train
        path: magahi/train-*
  - config_name: maithili
    data_files:
      - split: train
        path: maithili/train-*.parquet
  - config_name: manipuri
    data_files:
      - split: train
        path: manipuri/train-*.parquet
  - config_name: marwadi
    data_files:
      - split: train
        path: marwadi/train-*.parquet
  - config_name: nepali
    data_files:
      - split: train
        path: nepali/train-*.parquet
  - config_name: nimadi
    data_files:
      - split: train
        path: nimadi/train-*.parquet
  - config_name: odia
    data_files:
      - split: train
        path: odia/train-*.parquet
  - config_name: pahari_mahasui
    data_files:
      - split: train
        path: pahari_mahasui/train-*.parquet
  - config_name: rajasthani
    data_files:
      - split: train
        path: rajasthani/train-*
  - config_name: santali
    data_files:
      - split: train
        path: santali/train-*.parquet
  - config_name: sindhi
    data_files:
      - split: train
        path: sindhi/train-*.parquet
  - config_name: surgujia
    data_files:
      - split: train
        path: surgujia/train-*.parquet
  - config_name: tulu
    data_files:
      - split: train
        path: tulu/train-*.parquet
  - config_name: wagdi
    data_files:
      - split: train
        path: wagdi/train-*.parquet

Indic Dialect ASR Dataset

A multilingual ASR dataset covering 30 Indic dialect/languages with 2.8M+ samples.

Usage

from datasets import load_dataset

# Load a specific language
ds = load_dataset("grushaaaaa/indic-dialect-asr", "assamese", split="train")

Features

• audio: 16kHz WAV audio
• sentence: Transcription text
• language: Language name
• source: Source dataset