RNN及其變種LSTM/GRU/SRU

1. RNN h t = σ ( W ( h h ) h t − 1 + W ( h x ) x [ t ] ) (5) h_t=\sigma(W^{(hh)}h_{t-1}+W^{(hx)}x_{[t]}) \tag{5} ht​=σ(W(hh)ht−1​+W(hx)x[t]​)(5) y ^ t = s o f t m a x ( W ( S ) h t ) (6) \hat{y}_t=sof

>>阅读原文<<