BUILD.md

Build and Launch

In advance, you should clone the repository and enter directory by following the steps below.

git clone https://github.com/kekeho/gingerbreadman
cd gingerbreadman
git lfs pull

Server

Just three steps.

1. Build

docker-compose build  # CPU only
docker-compose -f docker-compose-gpu.yml build  # with Nvidia GPU

2. Set Password to Database

echo "DB_PASSWORD=password" > .env

3. Launch

docker-compose up  # CPU only
docker-compose -f docker-compose-gpu.yml up  # with Nvidia GPU

Front

1. Init

cd ./front
npm init

2. Build

os: "linux" or "mac" or "win"

npm run build:[os]

Worker

1. Build

docker-compose -f docker-compose-worker.yml build  # amd64
docker-compose -f docker-compose-worker-arm.yml build  # arm (tested on raspberry pi 3 64bit)

# arm with GPU (tested on jetson nano)
xargs -n 1 cp /usr/lib/aarch64-linux-gnu/libcudnn.so <<< "./services/face_location ./services/face_encoding"
docker-compose -f docker-compose-worker-arm-gpu.yml build

2. Config

Edit worker.env

NGINX_HOST=192.168.x.x
NGINX_PORT=8000

3. Launch

service: sex_detection or age_prediction or face_encoding or face_location

docker-compose -f docker-compose-worker.yml up  # amd64
docker-compose -f docker-compose-worker-arm.yml up  # arm (tested on raspberry pi 3 64bit)
docker-compose -f docker-compose-worker-arm-gpu.yml run [service]  # arm with GPU (tested on jetson nano) It is recommended to run only one container at a time.