Skip to content

Latest commit

 

History

History
23 lines (16 loc) · 1.22 KB

README.md

File metadata and controls

23 lines (16 loc) · 1.22 KB

Totem

A wrapper over Theano for quick creation and training of feedforward neural networks. Written as a part of the implementation for A Multi-scale Convolutional Neural Network Architecture for Music Auto-Tagging, the code for which can be found here. This was written in Python 2.

Requirements

  • Python 2.7
  • Theano >= 0.8
  • Numpy
  • Scikit-learn

Usage

The class model.Model represents a feedforward network. After initializing the model, the layers can be added to it using the model.add_layer function. The layers are defined in layers.py and contain most of the standard layers. Once all the layers have been added, the model.build function can be used to build the entire graph. The optimizer can be build using the model.build_optimizer method by passing one of the optimizers defined in optimizers.py.

A detailed working example is shown in MNIST_Example.py.

Installation

git clone https://github.com/many-facedgod/Totem
cd Totem
pip install .

Authors