Lam1360/YOLOv3-model-pruning
YOLOv3 object detection model with channel pruning (network slimming) for hand detection on the Oxford Hand dataset.
Not currently ranked — collecting fresh signals.
star history
This repository implements YOLOv3 for hand detection on the Oxford Hand dataset and applies channel pruning based on the Network Slimming paper (ICCV 2017). The approach uses sparse training on batch normalization gamma coefficients to identify and remove less important channels, reducing model parameters by 80%, FLOPs by 70%, while maintaining comparable mAP and doubling inference speed.
Frequently asked
- What is Lam1360/YOLOv3-model-pruning?
- YOLOv3 object detection model with channel pruning (network slimming) for hand detection on the Oxford Hand dataset.
- Is YOLOv3-model-pruning open source?
- Yes — Lam1360/YOLOv3-model-pruning is open source, released under the MIT license.
- What language is YOLOv3-model-pruning written in?
- Lam1360/YOLOv3-model-pruning is primarily written in Python.
- How popular is YOLOv3-model-pruning?
- Lam1360/YOLOv3-model-pruning has 1.7k stars on GitHub.
- Where can I find YOLOv3-model-pruning?
- Lam1360/YOLOv3-model-pruning is on GitHub at https://github.com/Lam1360/YOLOv3-model-pruning.