漫谈四种神经网络序列解码模型以及示例代码

for i, c in enumerate(vocab)) # 解码时须要将数字index映射成字符

inputs_train, tars_train = vectorize_stories(input_list, tar_list, word_to_idx, input_maxlen, tar_maxlen, vocab_size)

decoder_mode = 1 # 0 最简单模式，1 [1]向后模式，2 [2] Peek模式，3 [3]Attention模式

if decoder_mode == 3:

encoder_top_layer = LSTM(hidden_dim, return_sequences=True)

else:

encoder_top_layer = LSTM(hidden_dim)

if decoder_mode == 0:

decoder_top_layer = LSTM(hidden_dim, return_sequences=True)

decoder_top_layer.get_weights()

elif decoder_mode == 1:

decoder_top_layer = LSTMDecoder(hidden_dim=hidden_dim, output_dim=hidden_dim

, output_length=tar_maxlen, state_input=False, return_sequences=True)

elif decoder_mode == 2:

decoder_top_layer = LSTMDecoder2(hidden_dim=hidden_dim, output_dim=hidden_dim

, output_length=tar_maxlen, state_input=False, return_sequences=True)

elif decoder_mode == 3:

decoder_top_layer = AttentionDecoder(hidden_dim=hidden_dim, output_dim=hidden_dim

, output_length=tar_maxlen, state_input=False, return_sequences=True)

en_de_model = Sequential()

en_de_model.add(Embedding(input_dim=vocab_size,

output_dim=hidden_dim,

input_length=input_maxlen))

　　推荐阅读

　　如何使用batch-import工具向neo4j中导入海量数据

【引自T_SevenS的博客】在开辟neo4j的过程中，经常会有同窗问若何向neo4j中导入大年夜量的汗青数据，而这些数据一般都邑存在于关系型数据库中，如今本人就根据本身的导入经历，把导入的过>>>详细阅读

地址：http://www.17bianji.com/lsqh/34942.html