ASCEND / README.md
holylovenia's picture
Convert dataset to Parquet (#3)
737e980 verified
metadata
annotations_creators:
  - expert-generated
language_creators:
  - crowdsourced
language:
  - en
  - zh
license:
  - cc-by-sa-4.0
multilinguality:
  - multilingual
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - automatic-speech-recognition
task_ids: []
pretty_name: >-
  ASCEND: A Spontaneous Chinese-English Dataset for Code-switching in Multi-turn
  Conversation
tags:
  - speech-recognition
  - code-switching
dataset_info:
  config_name: main
  features:
    - name: id
      dtype: string
    - name: path
      dtype: string
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: transcription
      dtype: string
    - name: duration
      dtype: float32
    - name: language
      dtype: string
    - name: original_speaker_id
      dtype: int64
    - name: session_id
      dtype: int64
    - name: topic
      dtype: string
  splits:
    - name: train
      num_bytes: 1014573740.14
      num_examples: 9869
    - name: test
      num_bytes: 106171230.135
      num_examples: 1315
    - name: validation
      num_bytes: 106772517.43
      num_examples: 1130
  download_size: 1223536062
  dataset_size: 1227517487.7050002
configs:
  - config_name: main
    data_files:
      - split: train
        path: main/train-*
      - split: test
        path: main/test-*
      - split: validation
        path: main/validation-*
    default: true

Dataset Card for ASCEND

Table of Contents

Dataset Description

  • Homepage: [Needs More Information]
  • Repository: [Needs More Information]
  • Paper: https://arxiv.org/abs/2112.06223
  • Leaderboard: [Needs More Information]
  • Point of Contact: [Needs More Information]

Dataset Summary

ASCEND (A Spontaneous Chinese-English Dataset) introduces a high-quality resource of spontaneous multi-turn conversational dialogue Chinese-English code-switching corpus collected in Hong Kong. ASCEND consists of 10.62 hours of spontaneous speech with a total of ~12.3K utterances. The corpus is split into 3 sets: training, validation, and test with a ratio of 8:1:1 while maintaining a balanced gender proportion on each set.

Supported Tasks and Leaderboards

Code-switching

Languages

Chinese and English

Usage

To obtain the full dataset (complete with train, validation, and test set), simply run this:

import datasets
dataset = datasets.load_dataset("CAiRE/ASCEND")

Dataset Structure

A typical data point comprises the path to the audio file, the loaded audio array, and its transcription. Additional fields include datapoint id, duration, language, speaker id, session id, and topic.

{
    'id': '00644',
    'path': '.cache/huggingface/datasets/downloads/extracted/f0b33b5266cd9452ee310eef3577cf7adb7f29aa54dbff74b9a8ee406a55d614/waves/ses2_spk3_L13101_189.900_5.490.wav',
    'audio': {
        'path': '.cache/huggingface/datasets/downloads/extracted/f0b33b5266cd9452ee310eef3577cf7adb7f29aa54dbff74b9a8ee406a55d614/waves/ses2_spk3_L13101_189.900_5.490.wav',
        'array': array([-6.1035156e-05, -1.8310547e-04, 3.0517578e-05, ...,
            0.0000000e+00, -3.0517578e-05, 0.0000000e+00
        ], dtype = float32),
        'sampling_rate': 16000
    },
    'transcription': '因为你不可能邀你的female friends去说走我们去play basketball',
    'duration': 5.489999771118164,
    'language': 'mixed',
    'original_speaker_id': 3,
    'session_id': 2,
    'topic': 'sports'
}

Data Splits

Number of utterances: 9,869 train, 1,130 validation, and 1,315 test.

Additional Information

For comprehensive explanations, please check our paper.

Licensing Information

Creative Common Attribution Share-Alike 4.0 International (CC-BY-SA 4.0)

Citation Information

If you use our dataset, please cite us:

@inproceedings{lovenia2022ascend,
  title={ASCEND: A Spontaneous Chinese-English Dataset for Code-switching in Multi-turn Conversation},
  author={Lovenia, Holy and Cahyawijaya, Samuel and Winata, Genta Indra and Xu, Peng and Yan, Xu and Liu, Zihan and Frieske, Rita and Yu, Tiezheng and Dai, Wenliang and Barezi, Elham J and others},
  booktitle={Proceedings of the 13th Language Resources and Evaluation Conference (LREC)},
  year={2022}