yzd-v/FGD

Focal and Global Knowledge Distillation for Detectors (CVPR 2022)

44
/ 100
Emerging

This project helps computer vision practitioners improve the accuracy of their object detection models without needing to train much larger, more complex models from scratch. It takes an existing, pre-trained 'teacher' model and a 'student' model, and trains the student to learn from the teacher's expertise. The result is a smaller, more efficient object detection model that performs closer to a larger, more powerful one. Computer vision engineers and researchers who are working on optimizing object detection for real-world applications would use this.

385 stars. No commits in the last 6 months.

Use this if you need to deploy accurate object detection or instance segmentation models but are constrained by computational resources, and wish to improve a smaller model's performance by learning from a larger one.

Not ideal if you are starting a new object detection project and do not already have a trained 'teacher' model to distill knowledge from, or if you primarily need a general-purpose object detection framework.

object-detection computer-vision model-optimization deep-learning-deployment image-analysis
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

385

Forks

50

Language

Python

License

Apache-2.0

Last pushed

Sep 19, 2022

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/yzd-v/FGD"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.