README.md

Image Classifier Written with MXNet + Flask

A Flask (Python) Web Interface for MXNet Image Classifier.

This app simply invoked the pre-trained model provided by MXNet community.

Deployment Using Docker

docker run -p 8000:8000 xddeng/flask-app-for-mxnet-img-classifier:v2

Now you can try to access the service at http://localhost:8000

Deployment

Step - 1: Environment

sudo yum install python-pip
sudo yum install git
sudo yum install numpy opencv*

pip install Flask
pip install mxnet
pip install gunicorn

Step - 2: Clone This Project

git clone https://github.com/XD-DENG/flask-app-for-mxnet-img-classifier.git

Step - 3: Download Pre-Trained MXNet Model

From http://data.mxnet.io/models/imagenet-11k/, download

• resnet-152/resnet-152-symbol.json
• resnet-152/resnet-152-0000.params
• synset.txt

Note that we need to put all these three files under application directory.

Step - 4: Start Service

gunicorn -b 0.0.0.0:80 app:app