Bug in AttentionDecoderCell ? ?

[jjwangnlp]

Dear authors, I have three questions as follows:

Firstly,

seq2seq/seq2seq/cells.py

Line 87 in c020ccf

_xC = Lambda(lambda x: K.reshape(x, (-1, input_dim + hidden_dim)), output_shape=(input_dim + hidden_dim,))(_xC)

noted: output_shape=(input_dim + hidden_dim,) should it be ->-> output_shape=(input_length, input_dim + hidden_dim)???

Additionally, the same issue with #219 ,

instead of this:

seq2seq/seq2seq/cells.py

Line 85 in c020ccf

C = Lambda(lambda x: K.repeat(x, input_length), output_shape=(input_length, input_dim))(c_tm1)

shouldn't it be this (input_dim -> hidden_dim):

C = Lambda(lambda x: K.repeat(x, input_length), output_shape=(input_length, hidden_dim))(c_tm1)

Because c_tm1 has dimensionality of hidden_dim

Finally,

seq2seq/seq2seq/cells.py

Line 89 in c020ccf

alpha = W3(_xC)

line 89: alpha = W3(_xC) Here, did "Dense" play the same role as "TimedistributedDense" in keras? ?