VNCultureVQA: Visual Question Answering Dataset for Vietnamese Culture

VNCultureVQA is a Visual Question Answering dataset focused on Vietnamese culture, containing images with corresponding question–answer pairs. The dataset is divided into train and test sets based on difficulty levels. It was created by multimedia-synergy-lab and was last updated on 2026-05-07.

Use Cases

• Train Visual Question Answering models based on images and question-answer pairs about Vietnamese culture.
• Benchmark model performance on culturally-specific visual reasoning tasks based on difficulty-level splits.
• Fine-tune multimodal language models for applications in cultural heritage and education based on the described image-text pairs.

Strengths

• Dataset includes predefined train/test splits organized by difficulty level.
• Focus on Vietnamese culture provides a specific domain for model evaluation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

multimedia-synergy-lab on Hugging Face

Freshness

Last updated 2026-05-07 09:35:12; freshness should be verified.

Geography

Vietnam (implied by cultural focus)