andy-p / nngraph.jl

Deep Learning library for the Julia language

NNGraph.jl

NNGraph.jl is a Julia language package with functionality to construct arbitrary expression graphs over which the library can perform automatic differentiation similar to what you may find in Theano for Python, or in Torch etc. It is useful for creating arbitrary neural networks and was developed with a focus on enabling fast experimentation.

Example code

The core of the library is a Graph structure which maintains the links between matrices and how they are related th Here is how you would implement a simple Neural Network layer:

# T.B.D.
Will added soon ... I hope :)

To construct and train an LSTM for example, you would proceed as follows:

# T.B.D.
Will added soon ...

Credits

NNGraph.jl is based on a portion Andrej Karpathy's excellent RecurrentJS library in javascript. Speed enhancements were added by Iain Dunning. Some of Paul Heideman work on RecurrentNN.jl graph function have been used here.

License

MIT