Human Pose Estimation Model Serving on Multi-Model-Server (MMS)

Installation

1. Build the image for the MMS:

   $ docker build -t human-pose-mms .

2. Prepare the deployment model with model-archiver:

   $ docker run --rm -it -v $(pwd):/tmp human-pose-mms:latest bash
   
   $ model-archiver --model-name pose --model-path /tmp/service --handler handler:handle --runtime python3 --export-path /tmp

3. Start the MMS:

   $ docker run --rm -it -p 8080:8080 -p 8081:8081 -v $(pwd):/tmp mms-human-pose:latest multi-model-server --start --mms-config /tmp/config.properties --models posenet=pose.mar --model-store /tmp

Testing

Send a request with curl:

curl -X POST http://localhost:8080/predictions/posenet/ -T example.jpg

Running Client

1. Install python packages from here
2. Run the client

   $ python client/cli.py

Note

This repository is not for production use for sure!!!. Please read it on medium to learn more about this repository.

Thanks to the Joshua Earle for photo on Unsplash