Partial sharing of input and hidden-state weights in RNN, LSTM, and GRU yields about 50% parameter reduction with roughly unchanged perplexity on two language-modeling benchmarks.
Compressing recurrent neural network with tensor train,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Restricted Recurrent Neural Networks
Partial sharing of input and hidden-state weights in RNN, LSTM, and GRU yields about 50% parameter reduction with roughly unchanged perplexity on two language-modeling benchmarks.