leo-rain/lstm-matlab

LSTM-MATLAB is Long Short-term Memory (LSTM) in MATLAB, which is meant to be succinct, illustrative and for research purpose only. It is accompanied with a paper for reference: Revisit Long Short-Term Memory: An Optimization Perspective, NIPS deep learning workshop, 2014.

LSTM-MATLAB

LSTM-MATLAB is Long Short-term Memory (LSTM) in MATLAB, which is meant to be succinct, illustrative and for research purpose only. It is accompanied with a paper for reference: Revisit Long Short-Term Memory: An Optimization Perspective, NIPS deep learning workshop, 2014.

Creater & Maintainer Qi Lyu

#FEATURES

• original Long short-term Memory
• all connect peephole
• support optimization methods like LBFGS and CG
• CPU or GPU acceleration
• Mapreduce parallelization
• gradient checking
• easy configuration
• baseline experiment

#ACKNOWLEDGEMENTS The minFunc code folder included is provided by Mark Schmidt (http://www.cs.ubc.ca/~schmidtm). MATLAB Mapreduce is provided by Quoc V. Le(http://cs.stanford.edu/~quocle/optimizationWeb/index.html).

#USAGE To run the code, start from aStart.m. Data is generated by scripts in data directory on-the-fly. For faster LSTM implementation with complete features, see 'LSTMLayer' defined in C++ version. The dataset and labels etc follows the original LSTM paper in 1997.

License

MIT