liuwuliuyun/fast-object-detection-server

Fast Object Detection Server using YOLOv3

Installation and Running Guide for Fast Object Detection

Author: Yun Liu

Date: 2019/01/01

1. Devices, Drivers and System

1.1 GPU Devices

We only test this program on Nvidia GTX 1060 6GB and Nvidia GTX 1080 Ti, so make sure you have those devices before running this program.

1.2 Drivers

1.2.1 CUDA Drivers

Make sure you install CUDA 9.0 as your graphics card driver.

1.2.2 Cudnn

Make sure you install Cudnn 7.x on your computer.

1.3 System

We support Windows 10 x64 professional, but it should work fine with Ubuntu 16.04 LTS and Ubuntu 18.04 LTS.

2. Python Libs

Following packages should be installed before running:

PackageName Version

Click 7.0

Flask 1.0.2

Jinja2 2.10

MarkupSafe 1.1.0

Pillow 5.3.0

PyWavelets 1.0.1

Werkzeug 0.14.1

cloudpickle 0.6.1

cycler 0.10.0

dask 1.0.0

decorator 4.3.0

itsdangerous 1.1.0

kiwisolver 1.0.1

matplotlib 3.0.2

networkx 2.2

numpy 1.15.4

opencv-python 3.4.4.19

pip 10.0.1

pyparsing 2.3.0

python-dateutil 2.7.5

scikit-image 0.14.1

scipy 1.1.0

setuptools 39.0.1

six 1.11.0

toolz 0.9.0

torch 0.4.1

torchvision 0.2.1

tqdm 4.28.1

waitress 1.1.0

3.Running

3.1 Local

cd $ROOTDIR$ // cd into the root project directory

python app.py // start flask service and you can find this website running at http://localhost:5000/

3.2 Remote

You need to comment line 64 and uncomment line 66 in app.py first

cd $ROOTDIR$ // cd into the root project directory

python app.py // start flask service and you can find this website running at http://YOURIPADDR:8080/ and can be accessed remotely.

4.Trouble Shooting

First detection will be much slower because system need to load weights and networks into memory.