• Class tf.contrib.rnn.MultiRNNCell
• Class tf.nn.rnn_cell.MultiRNNCell

構建多隱層神經網絡

__init__(cells, state_is_tuple=True)

cells：rnn cell 的list

state_is_tuple：true，狀態Ct和ht就是分開記錄，放在一個tuple中，接受和返回的states是n-tuples，其中n=len(cells)，False，states是concatenated沿著列軸.後者即將棄用。

BasicLSTMCell 單隱層

BasicLSTMCell 多隱層

代碼示例

# encoding:utf-8
import tensorflow as tf

batch_size=10
depth=128

inputs=tf.Variable(tf.random_normal([batch_size,depth]))

previous_state0=(tf.random_normal([batch_size,100]),tf.random_normal([batch_size,100]))
previous_state1=(tf.random_normal([batch_size,200]),tf.random_normal([batch_size,200]))
previous_state2
 
=(tf.random_normal([batch_size,300]),tf.random_normal([batch_size,300]))

num_units=[100,200,300]
print(inputs)

cells=[tf.nn.rnn_cell.BasicLSTMCell(num_unit) for num_unit in num_units]
mul_cells=tf.nn.rnn_cell.MultiRNNCell(cells)

outputs,states=mul_cells(inputs,(previous_state0,previous_state1,previous_state2))

 
print(outputs.shape) #(10, 300)
print(states[0]) #第一層LSTM
print(states[1]) #第二層LSTM
print(states[2]) ##第三層LSTM
print(states[0].h.shape) #第一層LSTM的h狀態,(10, 100)
print(states[0].c.shape) #第一層LSTM的c狀態,(10, 100)
print(states[1].h.shape) #第二層LSTM的h狀態,(10, 200)

輸出

(10, 300)
LSTMStateTuple(c=<tf.Tensor ‘multi_rnn_cell/cell_0/basic_lstm_cell/Add_1:0‘ shape=(10, 100) dtype=float32>, h=<tf.Tensor ‘multi_rnn_cell/cell_0/basic_lstm_cell/Mul_2:0‘ shape=(10, 100) dtype=float32>)
LSTMStateTuple(c=<tf.Tensor ‘multi_rnn_cell/cell_1/basic_lstm_cell/Add_1:0‘ shape=(10, 200) dtype=float32>, h=<tf.Tensor ‘multi_rnn_cell/cell_1/basic_lstm_cell/Mul_2:0‘ shape=(10, 200) dtype=float32>)
LSTMStateTuple(c=<tf.Tensor ‘multi_rnn_cell/cell_2/basic_lstm_cell/Add_1:0‘ shape=(10, 300) dtype=float32>, h=<tf.Tensor ‘multi_rnn_cell/cell_2/basic_lstm_cell/Mul_2:0‘ shape=(10, 300) dtype=float32>)
(10, 100)
(10, 100)
(10, 200)

tf.nn.rnn_cell.MultiRNNCell