Can I read AnytimeYOLO: Optimizing Early Exits on EtoBox?

AnytimeYOLO: Optimizing Early Exits by Ngọc Ninh is a document available to read on EtoBox.

What is AnytimeYOLO: Optimizing Early Exits about?

The document introduces AnytimeYOLO, a family of YOLO architecture variants designed for anytime object detection, allowing for interruptible inference and early termination to provide predictions at any time. It presents a formal adaptation of anytime models, discusses a novel transposed YOLO variant, and proposes optimization algorithms for selecting optimal early-exit points. The paper evaluates the performance of these models on the MS COCO dataset and addresses deployment considerations for low-resourc

Author

Ngọc Ninh

Language

EN