An open source project from Data to AI Lab at MIT.

MLBazaar Demos

Jupyter Notebooks and Demos that to get started with the MLBazaar libraries.

Overview

This repository contains a collection of Jupyter Notebooks and Demos that will help you get started with the MLBazaar libraries:

• MLBlocks
• MLPrimitives
• BTB

The collection of notebooks is also available as a docker image in dockerhub.

Installation

The recommended way to run these demos is using mlbazaar/demos image available on https://hub.docker.com

Requirements

In order to run our docker image, the only requirement is to have Docker installed in your system.

Installation instructions for any possible system compatible can be found here

Running the Docker Image

Once you have docker installed, all you need to execute to run the image is:

docker run --rm -ti -p 8888:8888 mlbazaar/demos

Afterwards, you simply need to point your browser at http://127.0.0.1:8888 and start running the notebooks!

Running the demos locally

If you do not have Docker available, you can run the MLBazaar demos locally.

Requirements

The MLBazaar demos have been developed and tested on Python 3.6 using the dependencies listed inside the requirements.txt file.

Also, although it is not strictly required, the usage of a virtualenv is highly recommended in order to avoid interfering with other software installed in the system where the demos are run.

Setup

In order to run the demos, you will need to do the following steps:

1. Clone the repository
2. Create and activate a virtualenv using python3.6
3. Install the dependencies from the requirements.txt file.

git clone https://github.com/HDI-Project/mlprimitives-demos.git
cd mlprimitives-demos
pip install virtualenv
virtualenv -p $(which python3.6) venv
source mlprimitives-demos-venv/bin/activate
pip install -r requirements.txt

Running the notebook

Once everything is installed, you can start the jypter server with the following commands.

source mlprimitives-demos-venv/bin/activate
jupyter notebook