gistfile1.txt

def embedding_model():
    # define word embedding
    vocab_list = [(k, wv_model.wv[k]) for k, v in wv_model.wv.vocab.items()]
    embeddings_matrix = np.zeros((len(wv_model.wv.vocab.items()) + 1, wv_model.vector_size))
    for i in range(len(vocab_list)):
        word = vocab_list[i][0]
        embeddings_matrix[i + 1] = vocab_list[i][1]

    embedding_layer = Embedding(input_dim=len(embeddings_matrix),
                                output_dim=EMBEDDING_DIM,
                                weights=[embeddings_matrix],
                                trainable=False,name="Embedding")
    return embedding_layer,len(embeddings_matrix)

def ende_embedding_model(n_input, n_output, n_units):
    encoder_inputs = Input(shape=(None,), name="Encoder_input")

    encoder = LSTM(n_units,return_state=True, name='Encoder_lstm')
    Shared_Embedding,vocab_size = embedding_model()
    word_embedding_context = Shared_Embedding(encoder_inputs)
    encoder_outputs, state_h, state_c = encoder(word_embedding_context)
    encoder_states = [state_h, state_c]

    decoder_inputs = Input(shape=(None,), name="Decoder_input")
    decoder_lstm = LSTM(n_units, return_sequences=True, return_state=True, name="Decoder_lstm")
    word_embedding_answer = Shared_Embedding(decoder_inputs)
    decoder_outputs, _, _ = decoder_lstm(word_embedding_answer, initial_state=encoder_states)
    decoder_dense = TimeDistributed(Dense(vocab_size, activation='softmax', name="Dense_layer"))
    decoder_outputs = decoder_dense(decoder_outputs)
    model = Model([encoder_inputs, decoder_inputs], decoder_outputs)

    encoder_model = Model(encoder_inputs, encoder_states)
    decoder_state_input_h = Input(shape=(n_units,), name="H_state_input")
    decoder_state_input_c = Input(shape=(n_units,), name="C_state_input")
    decoder_states_inputs = [decoder_state_input_h, decoder_state_input_c]
    decoder_outputs, state_h, state_c = decoder_lstm(word_embedding_answer, initial_state=decoder_states_inputs)
    decoder_states = [state_h, state_c]
    decoder_outputs = decoder_dense(decoder_outputs)
    decoder_model = Model([decoder_inputs] + decoder_states_inputs, [decoder_outputs] + decoder_states)
    return model, encoder_model, decoder_model