FSU-QA: 850 Scene Question-Answer Pairs for Autonomous Driving

FSU-QA is a multimodal question-answering dataset for autonomous driving scenes. It contains 850 annotated scenes, with associated metadata and a fine-tuned Qwen3-VL-8B model. The dataset was created by Gong-Grant and last updated on Hugging Face in June 2026.

Use Cases

• Benchmarking visual question-answering models based on autonomous driving scenes.
• Fine-tuning large vision-language models for scene understanding tasks.
• Evaluating model performance on reasoning about spatial and temporal relationships in driving environments.

Strengths

• Contains 850 annotated scenes for question-answering.
• Includes a fine-tuned Qwen3-VL-8B model for direct application.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-06-25 15:18:11; freshness should be verified.

Provenance

Source

Hugging Face dataset repository by Gong-Grant.

Collection Method

Likely derived from nuScenes scene metadata and annotated with QA pairs.

Freshness

Last updated 2026-06-25 15:18:11.