SimonZeng7108/efficientsam3
EfficientSAM3 compresses SAM3 into lightweight, edge-friendly models via progressive knowledge distillation for fast promptable concept segmentation and tracking.
EfficientSAM3 helps you quickly identify and track specific objects or 'concepts' within images and videos, even on devices with limited computing power like mobile phones or embedded systems. You input an image or video and a text description of what you want to find, and it outputs a precise segmentation mask around that concept, or tracks its movement over time. This is ideal for field technicians, researchers, or anyone needing real-time object detection and tracking on edge devices.
449 stars.
Use this if you need to perform accurate, prompt-based object segmentation and tracking on images or videos using resource-constrained hardware such as drones, smart cameras, or mobile phones.
Not ideal if you primarily work with high-end server-side processing or cloud infrastructure where model size and real-time latency are not critical concerns.
Stars
449
Forks
35
Language
Jupyter Notebook
License
—
Category
Last pushed
Mar 07, 2026
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/computer-vision/SimonZeng7108/efficientsam3"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
uber-research/UPSNet
UPSNet: A Unified Panoptic Segmentation Network
augmentedstartups/AS-One
Easy & Modular Computer Vision Detectors, Trackers & SAM - Run YOLOv9,v8,v7,v6,v5,R,X in under...
xuebinqin/U-2-Net
The code for our newly accepted paper in Pattern Recognition 2020: "U^2-Net: Going Deeper with...
shreyas-bk/u2netdemo
Demonstration using Google Colab to show how U-2-NET can be used for Background Removal,...
robustsam/RobustSAM
RobustSAM: Segment Anything Robustly on Degraded Images (CVPR 2024 Highlight)