Code for paper "Full-Capacity Unitary Recurrent Neural Networks." Based on the complex_RNN repository from github.com/amarshah/complex_RNN.
Code coming soon for other experiments.
If you find this code useful, please cite the following references:
[1] M. Arjovsky, A. Shah, and Y. Bengio, “Unitary Evolution Recurrent Neural Networks,” Proc. International Conference on Machine Learning (ICML), 2016, pp. 1120–1128.
[2] S. Wisdom, T. Powers, J.R. Hershey, J. Le Roux, and L. Atlas, "Full-Capacity Unitary Recurrent Neural Networks," Advances in Neural Information Processing Systems (NIPS), 2016.
ToRun: run_their_LSTM.sh run_memory_lstms.sh
Todo: Start just cutting from the working bengio to give a minimalist version 20 mins Implement outs.