FVQA: Factual Visual Question Answering Dataset for Search-Augmented Training

FactualVQA (FVQA) is a multimodal Visual Question Answering dataset created for search-augmented training and evaluation. It was authored by lmms-lab and last updated on 2025-08-09. The dataset emphasizes knowledge-intensive questions that require external information beyond the given image.

Use Cases

• Training models for search-augmented visual question answering based on knowledge-intensive questions.
• Evaluating the ability of multimodal models to retrieve and integrate external factual information.
• Developing on-demand search strategies for AI systems based on image-question pairs.

Strengths

• Dataset is explicitly designed for search-augmented training and evaluation, a specific research focus.
• Each entry includes an image, a question, and an answer, providing a structured multimodal format.

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

lmms-lab

Freshness

Last updated 2025-08-09 11:39:47; freshness should be verified.