Table 1.

The main structures and parameters of DeepACLSTM

Layer Type	Size	NP
embedding	21	441
Convolution1D	1 × 42	1806
Convolution2D	3 × 1	168
Dropout1	0.5	0
FC 1	400	706,000
${\mathit{\text{LSTM}}}_f^1$	300	841,200
${\mathit{\text{LSTM}}}_b^1$	300	841,200
${\mathit{\text{LSTM}}}_f^2$	300	721,200
${\mathit{\text{LSTM}}}_b^2$	300	721,200
Dropout2	0.4	0
FC 2	600	600,600
Softmax	8	4808