Ioannis METAXAS

Published on Jul 25 · Submitted by Alberto Baldrati on Jul 28 · Samsung Research Abstract Large Vision-Language Models (LVLMs) remain bottlenecked by massive computational footprints, precluding their deployment on resource-constrained edge devices. While efforts to compress LVLMs focus heavily on vision token reduction or smaller language models, the vision encoder is largely overlooked, typically deployed as a monolithic, computationally heavy feature extractor.

1. Introduction Children’s physical activity is increasingly characterized by the specialization in a specific sport. Furthermore, the sport requirements and competition have been increased significantly as suggested by the high intensity of the workouts and the frequency of the competitions. With respect to rhythmic gymnastics and artistic gymnastics, the selection of the girls is made from a very young age.