yolo_1_pytorch

simplest implementation of yolo v1 via pytorch √

Language: 中文

paper: You Only Look Once: Unified, Real-Time Object Detection

CSDN blog: 博客解析

This repo is a brief implementation of yolo v1. You can easily train the model and visualize the result.

output tensor:	S×S×(B∗5+C)		S:	num of grids		B:	num of boxes		C:	num of classes
	7x7x(2*5+20)			7			2			20

---

git clone https://github.com/leviome/yolo_1_pytorch.git
cd yolo_1_pytorch

Environment:

• Python3
• Pytorch>=1.3
• cv2
• matplotlib

Dataset preparation

1. Download voc2007 dataset:

wget -c http://pjreddie.com/media/files/VOCtrainval_06-Nov-2007.tar
wget -c http://pjreddie.com/media/files/VOCtest_06-Nov-2007.tar
wget -c http://pjreddie.com/media/files/VOCdevkit_08-Jun-2007.tar

2. Extract all tars:

tar xvf VOCtrainval_06-Nov-2007.tar
tar xvf VOCtest_06-Nov-2007.tar
tar xvf VOCdevkit_08-Jun-2007.tar

3. put the data into dataset/voc2007 and make the folder structure look like:

dataset
├── voc2007
│   ├── Annotations
│   ├── ImageSets
│   ├── JPEGImages
│   ├── Label
│   ├── SegmentationClass
│   └── SegmentationObject
└── voc2012

4. fit voc dataset to yolo model as pytorch dataset format:

python fit_voc_to_yolo.py

Train

python train.py

Detect single image

python detect.py

Demo