metadata

language:
  - id
  - jv
  - su
dataset_info:
  features:
    - name: text
      dtype: string
    - name: url
      dtype: string
    - name: score
      dtype: float32
    - name: source
      dtype: string
  splits:
    - name: ace
      num_bytes: 221111651
      num_examples: 792594
    - name: ban
      num_bytes: 70299985
      num_examples: 244545
    - name: bjn
      num_bytes: 80927139
      num_examples: 296314
    - name: ind
      num_bytes: 2494248742
      num_examples: 8098063
    - name: jav
      num_bytes: 296261404
      num_examples: 1155142
    - name: min
      num_bytes: 20231649
      num_examples: 113323
    - name: sun
      num_bytes: 254676532
      num_examples: 894626
  download_size: 1936122223
  dataset_size: 3437757102
configs:
  - config_name: default
    data_files:
      - split: ace
        path: data/ace-*
      - split: ban
        path: data/ban-*
      - split: bjn
        path: data/bjn-*
      - split: ind
        path: data/ind-*
      - split: jav
        path: data/jav-*
      - split: min
        path: data/min-*
      - split: sun
        path: data/sun-*

Dataset Card for [acul3/KoPI-NLLB - NearDup Subset]

Dataset Summary

This dataset is a subset of the KoPI-NLLB, focusing only on near-duplicate entries across various Indonesian languages. It is provided in parquet format and includes multiple low-resource languages from Indonesia. The dataset features textual data with accompanying metadata (url, score, source), making it useful for natural language processing tasks such as translation, text similarity, and language modeling.

Languages

The dataset contains entries for the following Indonesian languages:

Acehnese (ace)
Balinese (ban)
Banjarese (bjn)
Indonesian (ind)
Javanese (jav)
Minangkabau (min)
Sundanese (sun)

Features

Each dataset contains the following features:

text: The text content in the respective language. url: The source URL where the text was extracted. score: A score indicating the similarity to a reference dataset. source: The originating source or dataset for the entry.

Usage

This dataset can be loaded using the following code snippet:

from datasets import load_dataset

dataset = load_dataset("acul3/KoPI-NLLB", name = "ace") # or "ban", "bjn", "ind", "jav", "min", "sun"

Citation

If you use this dataset, please cite the original KoPI-NLLB dataset and this subset. You can cite it as:

@inproceedings{acul3kopi2024,
  title={KoPI-NLLB: NearDup Subset},
  author={Acul3},
  year={2024},
  publisher={Hugging Face Datasets}
}