---
configs:
  - config_name: default
    data_files:
      - split: train
        path:
          - data/recitation_0/train/*.parquet
          - data/recitation_1/train/*.parquet
          - data/recitation_2/train/*.parquet
          - data/recitation_3/train/*.parquet
          - data/recitation_5/train/*.parquet
          - data/recitation_6/train/*.parquet
          - data/recitation_7/train/*.parquet
      - split: validation
        path:
          - data/recitation_0/validation/*.parquet
          - data/recitation_1/validation/*.parquet
          - data/recitation_2/validation/*.parquet
          - data/recitation_3/validation/*.parquet
          - data/recitation_5/validation/*.parquet
          - data/recitation_6/validation/*.parquet
          - data/recitation_7/validation/*.parquet
      - split: test
        path:
          - data/recitation_8/train/*.parquet
          - data/recitation_8/validation/*.parquet
dataset_info:
  splits:
    - name: train
      num_examples: 54823
    - name: test
      num_examples: 8787
    - name: validation
      num_examples: 7175

  featrues:
  - dtype: string
    name: aya_name
  - dtype: string
    name: aya_id
  - dtype: string
    name: reciter_name
  - dtype: int32
    name: recitation_id
  - dtype: string
    name: url
  - dtype:
      audio:
        decode: false
        sampling_rate: 16000
    name: audio
  - dtype: float32
    name: duration
  - dtype: float32
    name: speed
  - dtype:
      array2_d:
        dtype: float32
        shape:
        - null
        - 2
    name: speech_intervals
  - dtype: bool
    name: is_interval_complete
  - dtype: bool
    name: is_augmented
  - dtype:
      array2_d:
        dtype: float32
        shape:
        - null
        - 2
    name: input_features
  - dtype:
      array2_d:
        dtype: int32
        shape:
        - null
        - 1
    name: attention_mask
  - dtype:
      array2_d:
        dtype: int32
        shape:
        - null
        - 1
    name: labels
---
# Recitation Segmentations Dataset

This an modfied version of [this dataset](https://huggingface.co/datasets/obadx/recitation-segmentation) with these modifications:
* adding augmentation to the speed of the recitations utterance with column `speed` reflects the speed from 0.8 to 1.5 on 40% of the dataset using [audumentations](https://iver56.github.io/audiomentations/).
* adding data augmentation with [audiomentations](https://iver56.github.io/audiomentations/) on 40% of the dataset to prepare it for training the recitations spliter.

The codes for building this dataset is available at [github](https://github.com/obadx/recitations-segmenter)