adadelta_py

Clean & dependency-free implementation of the ADADELTA algorithm as presented in this paper

basic usage

The adadelta() function will need at least 2 things:

1. A gradient function that takes a vector and outputs the gradient of your objective function at that vector. (Vectors are just lists)
2. A initial guess to initialise the algorithm.

Given those 2, adadelta() will try minimise your function.

advanced usage

Check out the docstring to see all the other parameters that can be tweaked and further details.

Enjoy.