Study Maps Where Global Context and Weight Sharing Help Compact Vision Backbones
A new arXiv paper examines compact convolutional backbones like EfficientNet, which rely on local depthwise separable convolutions, and analyzes two structural limitations around global context and weight sharing.
Read the full story at arXiv cs.CV ↗
Timeline · 1 report
- 2026-11-19 05:00 · arXiv cs.CV
Where Global Context and Weight Sharing Work in Compact Vision Backbones