IndicVoices / README.md
eswardivi's picture
Upload dataset
4bcb1dc verified
|
raw
history blame
2.19 kB
metadata
dataset_info:
  - config_name: telugu_train
    features:
      - name: audio
        dtype: audio
      - name: end
        dtype: float64
      - name: text
        dtype: string
      - name: start
        dtype: float64
      - name: speaker_id
        dtype: int64
      - name: audio_raw_duration
        dtype: float64
      - name: scenario
        dtype: string
      - name: task_name
        dtype: string
    splits:
      - name: split1
        num_bytes: 1609730034
        num_examples: 1000
      - name: split2
        num_bytes: 1049782097
        num_examples: 1000
      - name: split3
        num_bytes: 1444386250
        num_examples: 1000
      - name: split4
        num_bytes: 3152422807
        num_examples: 1000
      - name: split5
        num_bytes: 5554666300
        num_examples: 1000
      - name: split6
        num_bytes: 4003049531
        num_examples: 500
      - name: split7
        num_bytes: 4053228693
        num_examples: 500
      - name: split8
        num_bytes: 1044856061
        num_examples: 344
      - name: split9
        num_bytes: 5059082700
        num_examples: 500
      - name: split10
        num_bytes: 4664957526
        num_examples: 500
    download_size: 29946910673
    dataset_size: 31636161999
  - config_name: telugu_valid
    features:
      - name: audio
        dtype: audio
      - name: text
        dtype: string
      - name: audio_raw_duration
        dtype: float64
      - name: scenario
        dtype: string
      - name: task_name
        dtype: string
    splits:
      - name: valid
        num_bytes: 314053800
        num_examples: 73
    download_size: 301914962
    dataset_size: 314053800
configs:
  - config_name: telugu_train
    data_files:
      - split: split1
        path: telugu_train/split1-*
      - split: split2
        path: telugu_train/split2-*
      - split: split3
        path: telugu_train/split3-*
      - split: split4
        path: telugu_train/split4-*
      - split: split5
        path: telugu_train/split5-*
      - split: split6
        path: telugu_train/split6-*
      - split: split7
        path: telugu_train/split7-*
      - split: split8
        path: telugu_train/split8-*
      - split: split9
        path: telugu_train/split9-*
      - split: split10
        path: telugu_train/split10-*
  - config_name: telugu_valid
    data_files:
      - split: valid
        path: telugu_valid/valid-*

This dataset is processed from (indicvoices)[https://ai4bharat.iitm.ac.in/indicvoices/]