Scikit-Luc

TLDR;

This is a python package containing the usefull script developed or found by Luc Giffon during his PhD thesis. Those scripts are mainly related to the machine learning and data processing field.

More informations

This package contains

• A simple API for downloading, loading and perform common operations on some datasets (MNIST and CIFAR10)
• Some kernel methods for approximation of phi mapping function (skluc.kernel)
• Some common operations on tensorflow neural networks (layers, training graph on Mnist)
• Some common example of various implementation tasks (mainly tensorflow stuff)

Installation from Pypi repository

Anywhere in your working virtualenv first install tensorflow using the documentation: https://www.tensorflow.org/install/ Then just do:

pip install scikit-luc

Installation from sources

In the root directory of the package, just do:

python setup.py install

or

pip install -e .

Documentation

See wiki of the git repository. https://gitlab.lis-lab.fr/luc.giffon/scikit-luc/wikis/Introduction

Convention

When some data is supposedly structured (e.g. images should be 2D array), the output of mldataset shouldn't take that into account by default.