Introduction

This Repository includes YOLOv3 with some lightweight backbones (ShuffleNetV2, MobileNet), some computer vision attention mechanism (SE Block, CBAM Block, ECA Block), pruning and quantization (reference to my senior's Repository https://github.com/coldlarry/YOLOv3-complete-pruning)

Environment

• python 3.7
• pytorch >= 1.1.0
• opencv-python

Datasets

• Oxfordhand datasets (1 class, including human's hand)
  https://pan.baidu.com/s/1ZYKXMEvNef41MdG1NgWYiQ (extract code: 00rw)
• Visdrone Remote datasets (10 classes, including pedestrian, car, bus, etc)
  https://pan.baidu.com/s/1JzJ6APRym8K64taZgcDZfQ (extract code: xyil)
• Bdd100K datasets (10 classes, including motor, train, traffic light, etc)
  https://pan.baidu.com/s/1dBrKEdy92Mxqg-JiyrVjkQ (extract code: lm4p)
• Dior datasets (20 classes, including airplane, airport, bridge, etc)
  https://pan.baidu.com/s/1Fc-zJtHy-6iIewvsKWPDnA (extract code: k2js)

Usage

1. Download the datasets, place them in the data directory
2. Train the models by using following command (change the model structure by changing the cfg file)

  python3 train.py --data data/oxfordhand.data --batch-size 16 --cfg cfg/yolov3-shufflenetv2-hand.cfg --img-size 608

3. Detect objects using the trained model (place the pictures or videos in the samples directory)

  python3 detect.py --cfg cfg/yolov3-shufflenetv2-hand.cfg --weights weights/backup100.pt --data data/oxfordhand.data

4. Results:
   
   
   

Changing YOLOv3 Backbone

ShuffleNetV2 + Two Scales Detection(YOLO Detector)

Using Oxfordhand datasets

Model	Params	Model Size	mAP
ShuffleNetV2 1x	3.57M	13.89MB	51.2
ShuffleNetV2 1.5x	5.07M	19.55MB	56.4
YOLOv3-tiny	8.67M	33.1MB	60.3

Using Visdrone datasets(Incomplete training)

Model	Params	Model Size	mAP
ShuffleNetV2 1x	3.59M	13.99MB	10.2
ShuffleNetV2 1.5x	5.09M	19.63MB	11
YOLOv3-tiny	8.69M	33.9MB	3.3

Attention Mechanism

Based on YOLOv3-tiny

SE Block paper : https://arxiv.org/abs/1709.01507
CBAM Block paper : https://arxiv.org/abs/1807.06521
ECA Block paper : https://arxiv.org/abs/1910.03151

Model	Params	mAP
YOLOv3-tiny	8.67M	60.3
YOLOv3-tiny + SE	8.933M	62.3
YOLOv3-tiny + CBAM	8.81M	62.7
YOLOv3-tiny + ECA	8.67M	62.6

Prune and Quantization

Based on my senior's Repository, reference to https://github.com/coldlarry/YOLOv3-complete-pruning

TODO

• ShuffleNetV2 backbone
• MobileNet backbone
• COCO datasets training
• Other detection strategies
• Other pruning strategies