ViOCRVQA: Vietnamese Text-in-Image Understanding Dataset

ViOCRVQA is a dataset for Vietnamese Optical Character Recognition and Visual Question Answering. It contains over 28,000 images and 120,000 question-answer pairs focused on text appearing in images. The dataset was created by huyhuy123 and was last updated on Hugging Face in February 2025.

Use Cases

• Train OCR models to read Vietnamese text from images based on the described image collection.
• Develop Visual Question Answering systems that answer questions about text present in images.
• Benchmark multimodal AI models on Vietnamese language understanding tasks.

Strengths

• Over 28,000 images provide a substantial visual corpus.
• More than 120,000 question-answer pairs offer extensive annotations for training.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

huyhuy123 on Hugging Face

Freshness

Last updated 2025-02-16 04:13:22; freshness should be verified.