GRU

Back to Recurrent Neural Networks

Gated Recurrent Unit. A simplified version of LSTM with fewer parameters: uses reset and update gates instead of three separate gates. Often performs comparably to LSTM with less computation.


deep-learning rnn gru