TinyGiantVLM: A Lightweight Vision-Language Architecture for Spatial Reasoning under Resource Constraints

Vinh-Thuan Ly, Hoang M. Truong, Xuan-Huong Nguyen

Reasoning about fine-grained spatial relationships in warehouse-scale environments poses a significant challenge for existing vision-language models (VLMs), which often struggle to…