ChavacanoMT / README.md
ajvicente's picture
Update README.md
a14d2e3 verified
metadata
license: cc-by-nc-4.0
language:
  - ceb
  - cbk
  - hil
  - tl
  - en
  - es
task_categories:
  - translation
pretty_name: ChavacanoMT Corpus
size_categories:
  - 100K<n<1M

ChavacanoMT Corpus

Description

ChavacanoMT is a benchmark corpus for the machine translation of Chavacano to and from related languages, Spanish, Cebuano, Hiligaynon, Tagalog, and English. The corpus consists of 767,053 parallel sentence samples from 15 language pairs. The corpus is a combination of multi-source and parallel sentences extracted from translations of the Bible and religious articles from the Jehovah's Witness website (with permission).

Usage

Each sentence pair in the text files is tab-delimited.

Limitations

The ChavacanoMT corpus primarily consists of religious texts due to the scarcity of parallel resources for this low-resource language. This composition may introduce stylistic and lexical biases, and some texts may not follow standardized orthography. Future versions will include more diverse and contemporary content to enhance linguistic coverage and model generalizability

Citation

@misc{chavacanomt_2024,
    title = "{ChavacanoMT: A Corpus and Evaluation of Neural Machine Translation for Philippine Creole Spanish [PREPRINT]}",
    author = {Aileen Joan Vicente and Theresse Faith Amamampang and Dunn Dexter Lahaylahay and Charibeth Cheng},
    year = "2024",
    address = "Research Square",
    publisher = "MIT Press",
    doi = "https://doi.org/10.21203/rs.3.rs-5022127/v1"

email [email protected] to request access