MogrifierLSTM
MogrifierLSTM copied to clipboard
RMichaelSwan
A quick walk-through of the innards of LSTMs and a naive implementation of the Mogrifier LSTM paper in PyTorch
I am a newbie in pytorch. It seems that the notebook use the self.Q and self.R for all rounds while the equation use Q_i and R_i for i th round?
Hello, there seem to be two differences from the paper: 1. In the paper, the Qi and Ri matrices are factorized as products of low-rank matrices, 2. and for different...
Does your code replicate the results in Mogrifier LSTM paper?
Metadata
Owner
Metadata
A quick walk-through of the innards of LSTMs and a naive implementation of the Mogrifier LSTM paper in PyTorch