https://jaypolra.github.io/tags/vision-language-models/Vision-Language ModelsHugo Blox Builder (https://hugoblox.com)en-usWed, 01 Jan 2025 00:00:00 +0000https://jaypolra.github.io/media/icon_hu_3943bc2ea2cd9fa1.pnghttps://jaypolra.github.io/tags/vision-language-models/https://jaypolra.github.io/project/vlm-hazard-reasoning/Wed, 01 Jan 2025 00:00:00 +0000https://jaypolra.github.io/project/vlm-hazard-reasoning/<p>YOLO can flag a red zone but cannot explain why the scene is dangerous. This project tests whether general-purpose VLMs can fill that explainability gap using structured domain-aware prompting.</p> <p>Prompts give models physical context about melt shop environments: what pot haulers are, what molten metal implies, what worker corridors mean for safety. Two-stage reasoning pipeline: first describe the scene neutrally, then evaluate against safety conditions.</p> <p>Output covers scene description, detected entities, spatial relationships, hazard assessment, risk level, and recommended action.</p>