README.md

metadata

pretty_name: SinMix2Mono
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    default: true
    data_files:
      - split: training_dataset
        path: training_dataset.csv
      - split: Golden_dataset
        path: Golden_dataset.csv
      - split: code_mixed_transliteration_ambiguity_dataset
        path: code-mixed_transliteration_ambiguity_dataset.csv
task_categories:
  - translation
language:
  - si
  - en
tags:
  - code-mixed
  - romanized Sinhala
  - transliteration

Dataset Card for SinMix2Mono

Dataset Summary

SinMix2Mono is a dataset of code-mixed text in Sinhala with ambiguity annotations and transliteration information.

Dataset Structure

The dataset contains three splits:

• training_dataset: Training data for model development
• Golden_dataset: High-quality golden standard for evaluation
• code-mixed_transliteration_ambiguity_dataset: Annotated data with code-mixing and transliteration ambiguity labels

Dataset Creation

The datasets are derived from the main SinMix2Mono collection.

Curation Rationale

Designed for research in code-mixed text processing, transliteration, and ambiguity resolution in Sinhala language.

Licensing Information

This dataset is licensed under the CC-BY-4.0 License.

Citation Information

// Citation would be added soon