pynet is a minimal Python library for dynamic automatic differentiation. The focus is on simplicity and it is meant to accompany the differentiation lecture of Deep Learning for Vision course.
pynet has a 2-clause BSD license. See file LICENSE for the complete license text.
pynet consists of Python code only. The directory structure is:
/pynet the pynet library
/pynet/net the core part of the library
/pynet/lib generic utilities
/examples concrete examples
pynet has been tested on Python 2.7.6. It does not have any requirements other than common Python libraries, in particular math, numpy, matplotlib, sys, os, string, copy, time, itertools, contextlib.
The usage of pynet is demonstrated through a toy machine learning example, given in two versions
python examples/spiral00.py
python examples/spiral.py -d
The first version is a single-file independent implementation that is not using the library, and the second is adapted with the use of pynet. Try -h for options.