# Gradient Harmonized Single-stage Detector

arXiv: https://arxiv.org/abs/1811.05181 · HTML: https://mlsanity.com/paper/1811.05181/

## Key facts

- Cited by 114 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Nov 2018.

## Built on by (most cited first, 30 of 114)

- [PV-RCNN++: Point-Voxel Feature Set Abstraction With Local Vector Representation for 3D Object Detection](https://mlsanity.com/paper/2102.00463/) (2021)
- [Efficient Sharpness-aware Minimization for Improved Training of Neural Networks](https://mlsanity.com/paper/2110.03141/) (2021)
- [Towards Large-Scale Small Object Detection: Survey and Benchmarks](https://mlsanity.com/paper/2207.14096/) (2022)
- [ResLT: Residual Learning for Long-tailed Recognition](https://mlsanity.com/paper/2101.10633/) (2021)
- [A Survey on Long-Tailed Visual Recognition](https://mlsanity.com/paper/2205.13775/) (2022)
- [RSG: A Simple but Effective Module for Learning Imbalanced Datasets](https://mlsanity.com/paper/2106.09859/) (2021)
- [FocalFormer3D : Focusing on Hard Instance for 3D Object Detection](https://mlsanity.com/paper/2308.04556/) (2023)
- [Hard Patches Mining for Masked Image Modeling](https://mlsanity.com/paper/2304.05919/) (2023)
- [PseCo: Pseudo Labeling and Consistency Training for Semi-Supervised Object Detection](https://mlsanity.com/paper/2203.16317/) (2022)
- [Deep Imbalanced Learning for Multimodal Emotion Recognition in Conversations](https://mlsanity.com/paper/2312.06337/) (2023)
- [Equalized Focal Loss for Dense Long-Tailed Object Detection](https://mlsanity.com/paper/2201.02593/) (2022)
- [Understanding the role of importance weighting for deep learning](https://mlsanity.com/paper/2103.15209/) (2021)
- [Towards Calibrated Model for Long-Tailed Visual Recognition from Prior Perspective](https://mlsanity.com/paper/2111.03874/) (2021)
- [DropLoss for Long-Tail Instance Segmentation](https://mlsanity.com/paper/2104.06402/) (2021)
- [Improving Contrastive Learning on Imbalanced Seed Data via Open-World Sampling](https://mlsanity.com/paper/2111.01004/) (2021)
- [Enhancing One-Shot Federated Learning Through Data and Ensemble Co-Boosting](https://mlsanity.com/paper/2402.15070/) (2024)
- [Single Model Deep Learning on Imbalanced Small Datasets for Skin Lesion Classification](https://mlsanity.com/paper/2102.01284/) (2021)
- [Balanced Knowledge Distillation for Long-tailed Learning](https://mlsanity.com/paper/2104.10510/) (2021)
- [OpenTAL: Towards Open Set Temporal Action Localization](https://mlsanity.com/paper/2203.05114/) (2022)
- [Resistance Training using Prior Bias: toward Unbiased Scene Graph Generation](https://mlsanity.com/paper/2201.06794/) (2022)
- [Seeking the Shape of Sound: An Adaptive Framework for Learning Voice-Face Association](https://mlsanity.com/paper/2103.07293/) (2021)
- [Loss Function Discovery for Object Detection via Convergence-Simulation Driven Search](https://mlsanity.com/paper/2102.04700/) (2021)
- [AutoLoss-Zero: Searching Loss Functions from Scratch for Generic Tasks](https://mlsanity.com/paper/2103.14026/) (2021)
- [Evaluation-oriented Knowledge Distillation for Deep Face Recognition](https://mlsanity.com/paper/2206.02325/) (2022)
- [Graph Classification by Mixture of Diverse Experts](https://mlsanity.com/paper/2103.15622/) (2021)
- [Scale-Equivalent Distillation for Semi-Supervised Object Detection](https://mlsanity.com/paper/2203.12244/) (2022)
- [Hard Sample Matters a Lot in Zero-Shot Quantization](https://mlsanity.com/paper/2303.13826/) (2023)
- [Dynamic Contrastive Distillation for Image-Text Retrieval](https://mlsanity.com/paper/2207.01426/) (2022)
- [Which Samples Should be Learned First: Easy or Hard?](https://mlsanity.com/paper/2110.05481/) (2021)
- [Reconcile Prediction Consistency for Balanced Object Detection](https://mlsanity.com/paper/2108.10809/) (2021)

## Abstract (arXiv)

Despite the great success of two-stage detectors, single-stage detector is still a more elegant and efficient way, yet suffers from the two well-known disharmonies during training, i.e. the huge difference in quantity between positive and negative examples as well as between easy and hard examples. In this work, we first point out that the essential effect of the two disharmonies can be summarized in term of the gradient. Further, we propose a novel gradient harmonizing mechanism (GHM) to be a hedging for the disharmonies. The philosophy behind GHM can be easily embedded into both classification loss function like cross-entropy (CE) and regression loss function like smooth-L1 (SL1) loss. To this end, two novel loss functions called GHM-C and GHM-R are designed to balancing the gradient flow for anchor classification and bounding box refinement, respectively. Ablation study on MS COCO demonstrates that without laborious hyper-parameter tuning, both GHM-C and GHM-R can bring substantial improvement for single-stage detector. Without any whistles and bells, our model achieves 41.6 mAP on COCO test-dev set which surpasses the state-of-the-art method, Focal Loss (FL) + SL1, by 0.8.