DyNet: The Dynamic Neural Network Toolkit

We describe DyNet, a toolkit for implementing neural network models based on dynamic declaration of network structure. In the static declaration strategy that is used in toolkits like Theano, CNTK, and TensorFlow, the user first defines a computation graph (a symbolic representation of the computation), and then examples are fed into an engine that executes this computation and computes its derivatives. In DyNet’s dynamic declaration strategy, computation graph construction is mostly transparent, being implicitly constructed by executing procedural code that computes the network outputs, and the user is free to use different network structures for each input. Dynamic declaration thus facilitates the implementation of more complicated network architectures, and DyNet is specifically designed to allow users to implement their models in a way that is idiomatic in their preferred programming language (C++ or Python). One challenge with dynamic declaration is that because the symbolic computation graph is defined anew for every training example, its construction must have low overhead. To achieve this, DyNet has an optimized C++ backend and lightweight graph representation. Experiments show that DyNet’s speeds are faster than or comparable with static declaration toolkits, and significantly faster than Chainer, another dynamic declaration toolkit. DyNet is released opensource under the Apache 2.0 license, and available at http://github.com/clab/dynet. Carnegie Mellon University, Pittsburgh, PA, USA Nara Institute of Science and Technology, Ikoma, Japan DeepMind, London, UK Bar Ilan University, Ramat Gan, Israel Allen Institute for Artificial Intelligence, Seattle, WA, USA University of Notre Dame, Notre Dame, IN, USA IBM T.J. Watson Research Center, Yorktown Heights, NY, USA University of Melbourne, Melbourne, Australia Johns Hopkins University, Baltimore, MD, USA Google, New York, NY, USA Google, Mountain View, CA, USA University of Washington, Seattle, USA Microsoft Research, Redmond, WA, USA University of Edinburgh, Edinburgh, UK 1 ar X iv :1 70 1. 03 98 0v 1 [ st at .M L ] 1 5 Ja n 20 17