Study Revisits Energy Use of Edge Vision-Language Models Under Continuous Workloads
A new arXiv paper argues that energy conclusions for edge vision-language models drawn from single-shot visual-prefill tests may not hold under continuous perception workloads, and examines optimization boundaries.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-03 05:00 · arXiv cs.CV
From Single-Shot QA to Continuous Perception: Revisiting the Energy Composition and Optimization Boundary of Edge VLMs