Training, Testing, and Packaging machine learning algorithms

Welcome to Deploy-ML

Deploy-ML is an open source, python package that simplifies the training, testing and deployment of machine learning algorithms.

To download, type in the following:

pip install deployml 

For documentation and roadmap, scroll down!

backdrop.jpg

SK-Learn

Screen Shot 2018-07-11 at 21.33.52.png

Basic tutorial covering the full flow of using Deploy-ML to train, test and deploy SK-Learn machine learning algorithms.

Documentation on the different types of SK-Learn machine learning algorithms supported by Deploy-ML

Documentation on training and testing functions supported for SK-Learn models.

Documentation on how to save/load your trained SK-Learn model with our easy to use deploy function. 


Keras: Basic neural network, training, testing, and deployment is now live. More advanced neural network structures are for the next release. Whilst this version supports pickling models for the NeuralNetworkBase it does not support deployment for the import model function. This will be addressed in the next version.

Screen Shot 2018-07-11 at 21.41.26.png

Basic tutorial covering the full flow of using Deploy-ML to train, test and deploy Keras machine learning algorithms.

Documentation on the different types of Keras machine learning algorithms supported by Deploy-ML

Documentation on training and testing functions supported for Keras models.

Documentation on how to save/load your trained Keras model with our easy to use deploy function. 


Flask

Screen Shot 2018-07-11 at 23.15.44.png

A big part of using machine learning is working out where to stick it and how to use it. Another goal for Deploy-ML is to develop a simple API tool for your saved machine learning algorithm. The Idea is to use Flask, as it is to be believed that this micro-framework could be leveraged to provide users a quick, easy and hassle free way of making a machine learning API. This is still in the ideas stage.


TensorFlow

Screen Shot 2018-07-11 at 21.39.42.png

Work supporting TensorFlow is underway, however, it's in it's infancy as we aim to get Keras support live before. 


JavaScript

Screen Shot 2018-07-12 at 10.40.44.png

Javascript support is still in the ideas phase. If possible, we aim to develop a loader that will enable javascript users to load the saved machine learning algorithms produced by Deploy-ML in their Javascript frame work enabling front-end and smartphone app utilisation. The aim is to have this in an NPM package.