Skip to content

A REST API to create and run Q-learning ML models using the Nupic framework.

Notifications You must be signed in to change notification settings

mellertson/nupic-predictor

Repository files navigation

Overview

This project implements an HTTP REST API as an interface to build and run machine learning models. The REST API is implemented using Flask. And the machine learning models are implemented using Nupic.

Because, this project's goal is to run machine learning models with real-time time series data I used data from Bitmex's Test Net.

NOTE

When launched on its on, this Docker container will only bring up a REST API interface, but will not instatiate nor run any machine learning models. To instantiate and run models you must make calls to the REST API.

Building The Docker Container

Because, Docker Compose is used you can build the container by executing the following in a terminal.

export VERSION=1.1 
VERSION=1.1 docker-compose build && \
docker push registry.cybertron.ninja/nupic_predictor:${VERSION}
unset VERSION

Deploying the Entire BAMM Stack on JONIN

There are three Docker stacks in the BAMM stack, which must be brought up in the following order:

  1. nupic-predictor
    • NOTE: This stack provisions the Docker networks for the entire BAMM stack.
  2. algo-backend
  3. crypto-trading-gym

To deploy the Nupic Predictor REST API executing the following on a Docker Swarm manager node.

docker stack deploy -c stack.jonin.yaml --with-registry-auth bamm-nupic-predictor

Once launched the REST API should be available at http://localhost:5000.

About

A REST API to create and run Q-learning ML models using the Nupic framework.

Resources

Stars

Watchers

Forks

Packages

No packages published