About the method of passing encoder states to decoder

[mrgloom]

Based on lstm_seq2seq.py example seems there is a better way to pass encoder states to decoder:

https://github.com/keras-team/keras/blob/7a39b6c62d43c25472b2c2476bd2a8983ae4f682/examples/lstm_seq2seq.py#L129-L132

https://github.com/keras-team/keras/blob/7a39b6c62d43c25472b2c2476bd2a8983ae4f682/examples/lstm_seq2seq.py#L139-L141

compared to

katakana/katakana/model.py

Lines 62 to 65 in 72d2f91

	encoder = LSTM(64, return_sequences=False)(encoder)
	decoder = Embedding(output_dict_size, 64, input_length=output_length, mask_zero=True)(decoder_input)
	decoder = LSTM(64, return_sequences=True)(decoder, initial_state=[encoder, encoder])