**Parago** is a command line tool that we use internally to rapidly build machine learning models, making it easy to customize for new use cases. Inspired by the likes of [Yeoman](, [Homebrew]( and even [NPM](https://www
Last updated a year ago by skafos .
Apache2 · Repository · Bugs · Original npm · Tarball · package.json
$ cnpm install parago 
SYNC missed versions from official npm registry.

Parago - Machine Learning Generator

Parago is a command line tool to help you rapidly create machine learning models by:

  • curating a list of community-submitted templates (called Generators)
  • abstracting common machine learning patterns and tasks into unified commands

Inspired by the likes of Yeoman, Homebrew and even NPM, Parago allows you to dive into machine learning by doing.

Why Parago?

Given the wide array of tools and technologies created to drive ML solutions, building & deploying models for apps should be simpler. At Skafos, we are tired of losing countless hours:

  • learning different training frameworks
  • fighting configuration and infrastructure
  • customizing every aspect of each new ML solution ...why start from scratch?

If you are like us, you would rather focus that time working on your apps. So we built Parago.


Install with npm:

$ npm install -g parago

Image Classification Example

Use Parago to create an image classifier that can identify cats and dogs. You can also skip right to the command & usage Documentation.

List Available Generators

$ pgo list

> Available Generators.
    - turicreate-image-classifier   : classify objects within an image

Create a new Project from a Generator

# Create project and enter the directory
$ pgo create myImageClassifer -g turicreate-image-classifier
$ cd myImageClassifier

Environment Setup

Environment setup is a hard problem to solve globally, especially when it comes to machine learning projects. Until we have a better solution (give us some ideas/feedback), we leave this up to the user to setup a virtual environment using Conda.

Each generator will have an environment.yml file included.

# Create and activate the conda Python environment
$ conda env create -f environment.yml
$ conda activate turicreate-image-classifier

Load Training Data

$ pgo data load --env data_src=cats_dogs

Train Model

$ pgo train --env epochs=30

Export to Core ML Format

$ pgo export --env output=coreml

For more details, checkout the full command & usage Documentation.

Submitting Your Own Generator

Parago is built on generators submitted by experts in the community. Each generator includes a set of commands, demonstrated above, and can be customized to meet the needs of your use-case. Want to submit a generator for others to pull and use? Great!

Clone the Generator Repo and submit a PR. Follow the structure of the examples listed.


Guidelines for contributions will also be made available this July. If you want to help, send us an email at

Core Tenets

At Skafos, we are a building tools for developers using community supported tools. We want to give back to the community that has helped us get to where we are. Since we released this under Apache2, we're hoping to work with YOU to develop a tool that we not only need, but believe will help others on the journey as well.

Our tenets:

  • Easy-to-use: Focus on building apps, not data science
  • Familiarity: No wasted hours learning a new tool
  • Learn-as-you-build: Ship apps while you learn
  • Extensive Library: Have many model generators to choose from
  • Ready To Deploy: Export to Core ML, TFLite and others formats to use in your apps

Get In Touch

We would love more involvement and are wide open to feedback, inspirations and ideas. Please throw a star or watch or even an issue on Parago, or throw out a tweet.

Feel free to email us at

Current Tags

  • 1.0.0                                ...           latest (a year ago)

5 Versions

  • 1.0.0                                ...           a year ago
  • 0.0.4                                ...           a year ago
  • 0.0.3                                ...           a year ago
  • 0.0.2                                ...           a year ago
  • 0.0.1                                ...           a year ago
Maintainers (1)
Today 0
This Week 0
This Month 0
Last Day 0
Last Week 5
Last Month 5
Dependencies (14)
Dev Dependencies (6)
Dependents (0)

Copyright 2014 - 2016 © |