Mapping Nematus model parameters to S2S

S2S Parameter	Value	Nematus Parameter	Explanation
dim-vocabs	<list>	n_words_src, n_words	Maximum items in vocabulary ordered by rank
dim-emb	int	dim_word	Size of embedding vector
dim-rnn	int	dim	Size of rnn hidden state
enc-type	str	NA	Type of encoder RNN : bidirectional, bi-unidirectional, alternating (s2s)
enc-cell	str	encoder	Type of RNN cell: gru, lstm, tanh (s2s)
enc-cell-depth	int	enc_recurrence_transition_depth	Number of tansitional cells in encoder layers (s2s)
enc-depth	int	enc_depth	Number of encoder layers (s2s)
dec-cell	str	decoder_deep	Type of RNN cell: gru, lstm, tanh (s2s)
dec-cell-base-depth	int	dec_base_recurrence_transition_depth	Number of tansitional cells in first decoder layer (s2s)
dec-cell-high-depth	int	dec_high_recurrence_transition_depth	Number of tansitional cells in next decoder layers (s2s)
dec-depth	int	dec_depth	Number of decoder layers (s2s)
skip		NA	Use skip connections (s2s)
layer-normalization	bool	layer_normalization	Enable layer normalization
best-deep		NA	Use WMT-2017-style deep configuration (s2s)
special-vocab		NA	Model-specific special vocabulary ids
tied-embeddings	bool	tie_decoder_embeddings	Tie target embeddings and output embeddings in output layer
tied-embeddings-src	bool		Tie source and target embeddings
tied-embeddings-all	bool		Tie all embedding layers and output layer