# EdgeSAM: Prompt-In-the-Loop Distillation for SAM

arXiv: https://arxiv.org/abs/2312.06660 · HTML: https://mlsanity.com/paper/2312.06660/

## Key facts

- Cited by 41 papers in the map (data through 14 Oct 2026).
- Builds on 57 of its 77 references that are in the map.
- First posted on arXiv 11 Dec 2023.

## Builds on (most cited first, 30 of 57)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (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)
- [MobileNetV2: Inverted Residuals and Linear Bottlenecks](https://mlsanity.com/paper/1801.04381/) (2018)
- [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://mlsanity.com/paper/1704.04861/) (2017)
- [Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains](https://mlsanity.com/paper/2006.10739/) (2020)
- [Searching for MobileNetV3](https://mlsanity.com/paper/1905.02244/) (2019)
- [MMDetection: Open MMLab Detection Toolbox and Benchmark](https://mlsanity.com/paper/1906.07155/) (2019)
- [Self-training with Noisy Student improves ImageNet classification](https://mlsanity.com/paper/1911.04252/) (2019)
- [Cascade R-CNN: Delving into High Quality Object Detection](https://mlsanity.com/paper/1712.00726/) (2017)
- [LVIS: A Dataset for Large Vocabulary Instance Segmentation](https://mlsanity.com/paper/1908.03195/) (2019)
- [ShuffleNet: An Extremely Efficient Convolutional Neural Network for Mobile Devices](https://mlsanity.com/paper/1707.01083/) (2017)
- [ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design](https://mlsanity.com/paper/1807.11164/) (2018)
- [SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size](https://mlsanity.com/paper/1602.07360/) (2016)
- [EfficientDet: Scalable and Efficient Object Detection](https://mlsanity.com/paper/1911.09070/) (2019)
- [Exploring Plain Vision Transformer Backbones for Object Detection](https://mlsanity.com/paper/2203.16527/) (2022)
- [BiSeNet: Bilateral Segmentation Network for Real-time Semantic Segmentation](https://mlsanity.com/paper/1808.00897/) (2018)
- [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://mlsanity.com/paper/2110.02178/) (2021)
- [Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations](https://mlsanity.com/paper/1707.03237/) (2017)
- [Detecting Twenty-thousand Classes using Image-level Supervision](https://mlsanity.com/paper/2201.02605/) (2022)
- [YOLACT: Real-time Instance Segmentation](https://mlsanity.com/paper/1904.02689/) (2019)
- [GhostNet: More Features from Cheap Operations](https://mlsanity.com/paper/1911.11907/) (2019)
- [Faster Segment Anything: Towards Lightweight SAM for Mobile Applications](https://mlsanity.com/paper/2306.14289/) (2023)
- [BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation](https://mlsanity.com/paper/2004.02147/) (2020)
- [Structured Knowledge Distillation for Dense Prediction](https://mlsanity.com/paper/1903.04197/) (2019)
- [ICNet for Real-Time Semantic Segmentation on High-Resolution Images](https://mlsanity.com/paper/1704.08545/) (2017)
- [Mobile-Former: Bridging MobileNet and Transformer](https://mlsanity.com/paper/2108.05895/) (2021)

## Built on by (most cited first, 30 of 41)

- [Osprey: Pixel Understanding with Visual Instruction Tuning](https://mlsanity.com/paper/2312.10032/) (2023)
- [Open-Vocabulary SAM: Segment and Recognize Twenty-thousand Classes Interactively](https://mlsanity.com/paper/2401.02955/) (2024)
- [Towards Language-Driven Video Inpainting via Multimodal Large Language Models](https://mlsanity.com/paper/2401.10226/) (2024)
- [Hi-SAM: Marrying Segment Anything Model for Hierarchical Text Segmentation](https://mlsanity.com/paper/2401.17904/) (2024)
- [Fast-FoundationStereo: Real-Time Zero-Shot Stereo Matching](https://mlsanity.com/paper/2512.11130/) (2025)
- [MatAnyone: Stable Video Matting with Consistent Memory Propagation](https://mlsanity.com/paper/2501.14677/) (2025)
- [MG-LLaVA: Towards Multi-Granularity Visual Instruction Tuning](https://mlsanity.com/paper/2406.17770/) (2024)
- [EdgeTAM: On-Device Track Anything Model](https://mlsanity.com/paper/2501.07256/) (2025)
- [Mamba or RWKV: Exploring High-Quality and High-Efficiency Segment Anything Model](https://mlsanity.com/paper/2406.19369/) (2024)
- [MedSAM-U: Uncertainty-Guided Auto Multi-Prompt Adaptation for Reliable MedSAM](https://mlsanity.com/paper/2409.00924/) (2024)
- [BA-SAM: Scalable Bias-Mode Attention Mask for Segment Anything Model](https://mlsanity.com/paper/2401.02317/) (2024)
- [Explore In-Context Segmentation via Latent Diffusion Models](https://mlsanity.com/paper/2403.09616/) (2024)
- [De-LightSAM: Modality-Decoupled Lightweight SAM for Generalizable Medical Segmentation](https://mlsanity.com/paper/2407.14153/) (2024)
- [Swiss Army Knife: Synergizing Biases in Knowledge from Vision Foundation Models for Multi-Task Learning](https://mlsanity.com/paper/2410.14633/) (2024)
- [RMP-SAM: Towards Real-Time Multi-Purpose Segment Anything](https://mlsanity.com/paper/2401.10228/) (2024)
- [Unleashing the Power of Generic Segmentation Models: A Simple Baseline for Infrared Small Target Detection](https://mlsanity.com/paper/2409.04714/) (2024)
- [DifferSeg: Towards Diverse Multimodal Binary Segmentation via Differential Perception and Frequency Guidance](https://mlsanity.com/paper/2606.08906/) (2026)
- [Lite-SAM Is Actually What You Need for Segment Everything](https://mlsanity.com/paper/2407.08965/) (2024)
- [EMOv2: Pushing 5M Vision Model Frontier](https://mlsanity.com/paper/2412.06674/) (2024)
- [Segment This Thing: Foveated Tokenization for Efficient Point-Prompted Segmentation](https://mlsanity.com/paper/2506.11131/) (2025)
- [Hierarchical Self-Prompting SAM: A Prompt-Free Medical Image Segmentation Framework](https://mlsanity.com/paper/2506.02854/) (2025)
- [EfficientViT-SAM: Accelerated Segment Anything Model Without Accuracy Loss](https://mlsanity.com/paper/2402.05008/) (2024)
- [AHCQ-SAM: Toward Accurate and Hardware-Compatible Post-Training Segment Anything Model Quantization](https://mlsanity.com/paper/2503.03088/) (2025)
- [Agglomerating Large Vision Encoders via Distillation for VFSS Segmentation](https://mlsanity.com/paper/2504.02351/) (2025)
- [Prompt Engineering in Segment Anything Model: Methodologies, Applications, and Emerging Challenges](https://mlsanity.com/paper/2507.09562/) (2025)
- [Words to Wheels: Vision-Based Autonomous Driving Understanding Human Language Instructions Using Foundation Models](https://mlsanity.com/paper/2410.10577/) (2024)
- [Compress Any Segment Anything Model (SAM)](https://mlsanity.com/paper/2507.08765/) (2025)
- [The SAM2-to-SAM3 Gap in the Segment Anything Model Family: Why Prompt-Based Expertise Fails in Concept-Driven Image Segmentation](https://mlsanity.com/paper/2512.06032/) (2025)
- [VCA: Vision-Click-Action Framework for Precise Manipulation of Segmented Objects in Target Ambiguous Environments](https://mlsanity.com/paper/2602.23583/) (2026)
- [StructSAM: Structure- and Spectrum-Preserving Token Merging for Segment Anything Models](https://mlsanity.com/paper/2603.07307/) (2026)

## Abstract (arXiv)

This paper presents EdgeSAM, an accelerated variant of the Segment Anything Model (SAM), optimized for efficient execution on edge devices with minimal compromise in performance. Our approach involves distilling the original ViT-based SAM image encoder into a purely CNN-based architecture, better suited for edge devices. We carefully benchmark various distillation strategies and demonstrate that task-agnostic encoder distillation fails to capture the full knowledge embodied in SAM. To overcome this bottleneck, we include both the prompt encoder and mask decoder in the distillation process, with box and point prompts in the loop, so that the distilled model can accurately capture the intricate dynamics between user input and mask generation. To mitigate dataset bias issues stemming from point prompt distillation, we incorporate a lightweight module within the encoder. As a result, EdgeSAM achieves a 37-fold speed increase compared to the original SAM, and it also outperforms MobileSAM/EfficientSAM, being over 7 times as fast when deployed on edge devices while enhancing the mIoUs on COCO and LVIS by 2.3/1.5 and 3.1/1.6, respectively. It is also the first SAM variant that can run at over 30 FPS on an iPhone 14. Code and demo are available at https://www.mmlab-ntu.com/project/edgesam.