ViDia2Std: A Parallel Corpus for Vietnamese Dialect-to-Standard Translation

ViDia2Std is a high-quality, manually annotated parallel corpus for translating Vietnamese dialects (Northern, Central, Southern) into Standard Vietnamese. Sourced from authentic social media comments across all 63 provinces of Vietnam, it serves as a benchmark for dialect normalization and adaptation tasks.

This dataset accompanies the paper: "ViDia2Std: A Parallel Corpus and Methods for Low-Resource Vietnamese Dialect-to-Standard Translation" (Accepted at AAAI-26).

📂 Dataset Structure

The dataset is split into three files:

• train.csv: The training set (10,870 pairs).
• dev.csv: The development/validation set (1,184 pairs).
• test.csv: The testing set (1,603 pairs).

📝 Data Fields

Column	Description
dialect	The original dialectal input sentence (source). Contains regional words, and non-standard grammar.
standard	The manually normalized Standard Vietnamese sentence (target). Equivalent in meaning and intent to the source.
region	The dialect region of the source sentence: northern, central, or southern.
sentiment	(Only in test.csv) The sentiment label of the sentence (positive, negative, neutral). Used for extrinsic evaluation.

📊 Statistics

Split	Samples
Train	10,870
Dev	1,184
Test	1,603
Total	13,657

🚀 Usage

Loading with Pandas

import pandas as pd

train_df = pd.read_csv("https://huggingface.co/datasets/Biu3010/ViDia2Std/resolve/main/train.csv").dropna()
valid_df = pd.read_csv("https://huggingface.co/datasets/Biu3010/ViDia2Std/resolve/main/dev.csv").dropna()
test_df  = pd.read_csv("https://huggingface.co/datasets/Biu3010/ViDia2Std/resolve/main/test.csv").dropna()

print(train_df.head())

📝 Citation

If you use this dataset, please cite our publication ViDia2Std: A Parallel Corpus and Methods for Low-Resource Vietnamese Dialect-to-Standard Translation:

@article{Anh_Ta_Van_Dinh_Nguyen_2026,
  title={ViDia2Std: A Parallel Corpus and Methods for Low-Resource Vietnamese Dialect-to-Standard Translation},
  volume={40},
  url={https://ojs.aaai.org/index.php/AAAI/article/view/40247},
  DOI={10.1609/aaai.v40i36.40247},
  abstractNote={Vietnamese exhibits extensive dialectal variation, posing challenges for NLP systems trained predominantly on standard Vietnamese. Such systems often underperform on dialectal inputs, especially from underrepresented Central and Southern regions. Previous work on dialect normalization has focused narrowly on Central-to-Northern dialect transfer using synthetic data and limited dialectal diversity. These efforts exclude Southern varieties and intra-regional variants within the North. We introduce ViDia2Std, the first manually annotated parallel corpus for dialect-to-standard Vietnamese translation covering all 63 provinces. Unlike prior datasets, ViDia2Std includes diverse dialects from Central, Southern, and non-standard Northern regions often absent from existing resources, making it the most dialectally inclusive corpus to date. The dataset consists of over 13,000 sentence pairs sourced from real-world Facebook comments and annotated by native speakers across all three dialect regions. To assess annotation consistency, we define a semantic mapping agreement metric that accounts for synonymous standard mappings across annotators. Based on this criterion, we report agreement rates of 86% (North), 82% (Central), and 85% (South). We benchmark several sequence-to-sequence models on ViDia2Std. mBART-large-50 achieves the best results (BLEU 0.8166, ROUGE-L 0.9384, METEOR 0.8925), while ViT5-base offers competitive performance with fewer parameters. ViDia2Std demonstrates that dialect normalization substantially improves downstream tasks, highlighting the need for dialect-aware resources in building robust Vietnamese NLP systems.},
  number={36},
  journal={Proceedings of the AAAI Conference on Artificial Intelligence},
  author={Anh Ta, Khoa and Van Dinh, Nguyen and Nguyen, Kiet Van},
  year={2026},
  month={Mar.},
  pages={29995-30004}
}