caffe has become very popular as a deep learning framework. However, its prototxt files can be hundreds of lines long, and attempting to define models using pycaffe can be frustrating. rambutan is a python wrapper for caffe which aims at providing a simple, pythonic, interface for users so that users can define, train, and evaluate deep models in only a few lines of code. It requires that caffe and pycaffe are both built properly.
Installation is as simple as pip install rambutan to get the latest released version, or the following to get the bleeding edge:
git clone https://github.com/jmschrei/rambutan.git
cd rambutan
python setup.py install