Data Flow Facilitator for Machine Learning (dffml)

DFFML provides APIs for dataset generation and storage, and model definition using any machine learning framework, from high level down to low level use is supported.

The goal of DFFML is to build a community driven library of plugins for dataset generation and model definition. So that we as developers and researchers can quickly and easily plug and play various pieces of data with various model implementations.

Here's a quick demo showing how DFFML can be used to train on the iris dataset. The more we build up the library of plugins (which anyone can maintain, they don't have to be contributed upstream unless you want to) the more variations on model implementations and feature data generators we all have to work with.

Right now we've released a wrapper around the Tensorflow DNN estimator, and a set of feature generators which gather data from git repositories.

Installation

DFFML currently should work with Python 3.6. However, only Python 3.7 is officially supported. This is because there are a lot of nice helper methods Python 3.7 implemented that we intend to use instead of re-implementing.

python3.7 -m pip install -U dffml

You can also install the Features for Git Version Control, and Models for Tensorflow Library all at once.

• DFFML Features for Git Version Control
• DFFML Models for Tensorflow Library

If you want a quick how to on the iris dataset head to the DFFML Models for Tensorflow Library repo.

python3.7 -m pip install -U dffml[git,tensorflow]

If you don't have Python 3.7 we have a docker image for you, or you can install pyenv which will quickly and easily give you Python 3.7. See docs/INSTALL.md for more details.

Usage

To start using dffml for data set generation with a single CLI command see DFFML Features for Git Version Control.

To start using dffml for machine learning with a few CLI commands see DFFML Models for Tensorflow Library.

Documentation

Start with Architecture.

Contributing

DFFML is meant to be a community driven application. There are various segments you can help with:

• Found a bug, error in the docs or have a new idea, create an issue here.
• Help us fix an issue, refer to HACKING in the docs.
• Make sure you go through CONTRIBUTING.md before contributing.

Tutorials

Got an idea for a new feature/model, tutorials will help you write code that takes full advantage of the DFFML API. Making your next machine learning project a breeze to write!

• Features: The new feature tutorial will walk you through how to write a new DFFML feature to generate new data for a dataset.
• Models: The new model tutorial will walk you through how to wrap your favorite framework or a custom implementation in the DFFML library's model API.

License

dffml is distributed under the MIT License.

Legal

This software is subject to the U.S. Export Administration Regulations and other U.S. law, and may not be exported or re-exported to certain countries (Cuba, Iran, Crimea Region of Ukraine, North Korea, Sudan, and Syria) or to persons or entities prohibited from receiving U.S. exports (including Denied Parties, Specially Designated Nationals, and entities on the Bureau of Export Administration Entity List or involved with missile technology or nuclear, chemical or biological weapons).