# AnyDepth-DETR/-YOLO: Any-depth object detection with a single network

arXiv: https://arxiv.org/abs/2605.09407 · HTML: https://mlsanity.com/paper/2605.09407/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 28 of its 40 references that are in the map.
- First posted on arXiv 10 May 2026.

## Builds on (most cited first, 28 of 28)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Feature Pyramid Networks for Object Detection](https://mlsanity.com/paper/1612.03144/) (2016)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [YOLOv4: Optimal Speed and Accuracy of Object Detection](https://mlsanity.com/paper/2004.10934/) (2020)
- [DINO: DETR with Improved DeNoising Anchor Boxes for End-to-End Object Detection](https://mlsanity.com/paper/2203.03605/) (2022)
- [EfficientDet: Scalable and Efficient Object Detection](https://mlsanity.com/paper/1911.09070/) (2019)
- [YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors](https://mlsanity.com/paper/2207.02696/) (2022)
- [Path Aggregation Network for Instance Segmentation](https://mlsanity.com/paper/1803.01534/) (2018)
- [DAB-DETR: Dynamic Anchor Boxes are Better Queries for DETR](https://mlsanity.com/paper/2201.12329/) (2022)
- [DETRs Beat YOLOs on Real-time Object Detection](https://mlsanity.com/paper/2304.08069/) (2023)
- [DN-DETR: Accelerate DETR Training by Introducing Query DeNoising](https://mlsanity.com/paper/2203.01305/) (2022)
- [YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information](https://mlsanity.com/paper/2402.13616/) (2024)
- [CSPNet: A New Backbone that can Enhance Learning Capability of CNN](https://mlsanity.com/paper/1911.11929/) (2019)
- [Slimmable Neural Networks](https://mlsanity.com/paper/1812.08928/) (2018)
- [YOLOv12: Attention-Centric Real-Time Object Detectors](https://mlsanity.com/paper/2502.12524/) (2025)
- [Billion-scale semi-supervised learning for image classification](https://mlsanity.com/paper/1905.00546/) (2019)
- [Scaled-YOLOv4: Scaling Cross Stage Partial Network](https://mlsanity.com/paper/2011.08036/) (2020)
- [Focal and Global Knowledge Distillation for Detectors](https://mlsanity.com/paper/2111.11837/) (2021)
- [General Instance Distillation for Object Detection](https://mlsanity.com/paper/2103.02340/) (2021)
- [Localization Distillation for Dense Object Detection](https://mlsanity.com/paper/2102.12252/) (2021)
- [CrossKD: Cross-Head Knowledge Distillation for Object Detection](https://mlsanity.com/paper/2306.11369/) (2023)
- [PKD: General Distillation Framework for Object Detectors via Pearson Correlation Coefficient](https://mlsanity.com/paper/2207.02039/) (2022)
- [DynamicDet: A Unified Dynamic Architecture for Object Detection](https://mlsanity.com/paper/2304.05552/) (2023)
- [KD-DETR: Knowledge Distillation for Detection Transformer with Consistent Distillation Points Sampling](https://mlsanity.com/paper/2211.08071/) (2022)
- [EAutoDet: Efficient Architecture Search for Object Detection](https://mlsanity.com/paper/2203.10747/) (2022)
- [You Only Look Once at Anytime (AnytimeYOLO): Analysis and Optimization of Early-Exits for Object-Detection](https://mlsanity.com/paper/2503.17497/) (2025)
- [Adaptive Depth Networks with Skippable Sub-Paths](https://mlsanity.com/paper/2312.16392/) (2023)

## Abstract (arXiv)

Modern object detectors are static, fixed-depth networks optimized for a single operating point, requiring separate models for different deployment scenarios. We present an any-depth detection framework that enables a single network to span a continuous range of accuracy--efficiency trade-offs by controlling depth at inference time without retraining. Each backbone and neck stage is divided into an essential path, which always executes, and a skippable refinement path; this decomposition preserves the full multi-scale feature hierarchy at every depth configuration, unlike conventional early exiting that discards entire stages. To train such a network, jointly optimizing many sub-networks of varying depth introduces conflicting gradient signals. We address this via self-distillation between only the two extremes, with prediction-level and feature-level alignment losses that enforce stage-wise modularity, ensuring the outputs of each stage remain compatible regardless of the paths taken. Instantiated on RT-DETR and YOLOv12, our full-depth configurations match or surpass their respective SOTA baselines with negligible parameter overhead, while the most efficient configurations achieve up to 1.82× speedup at a cost of only 2.0 AP, all from a single set of weights.