我正在尝试建立一个用于文本生成的编码器-解码器模型。我正在使用带有嵌入层的LSTM层。我在从嵌入层到LSTM编码器层的输出中遇到了一个问题。我得到的错误是：

ValueError: Input 0 of layer lstm is incompatible with the layer: expected ndim=3, found ndim=4. Full shape received: (None, 13, 128, 512)

我的编码器数据具有形状：(40, 13, 128) = (num_observations, max_encoder_seq_length, vocab_size) ×嵌入大小/潜在尺寸= 512。
我的问题是：我怎样才能"摆脱"从嵌入层到LSTM编码器层的第四维，或者换句话说：我应该如何将这4个维度传递到编码器模型的LSTM层？2由于我是这个主题的新手，我最终还应该在解码器LSTM层中纠正什么？
我读过一些帖子，包括this，这个one和许多其他的，但没有找到一个解决方案。在我看来，我的问题不是在模型中，而是在数据的形状。任何提示或评论可能是错误的将是非常感谢。非常感谢
我的模型如下所示，来自（this tutorial）：

encoder_inputs = Input(shape=(max_encoder_seq_length,))
x = Embedding(num_encoder_tokens, latent_dim)(encoder_inputs)
x, state_h, state_c = LSTM(latent_dim, return_state=True)(x)
encoder_states = [state_h, state_c]

# Set up the decoder, using `encoder_states` as initial state.
decoder_inputs = Input(shape=(max_decoder_seq_length,))
x = Embedding(num_decoder_tokens, latent_dim)(decoder_inputs)
x = LSTM(latent_dim, return_sequences=True)(x, initial_state=encoder_states)
decoder_outputs = Dense(num_decoder_tokens, activation='softmax')(x)

# Define the model that will turn
# `encoder_input_data` & `decoder_input_data` into `decoder_target_data`
model = Model([encoder_inputs, decoder_inputs], decoder_outputs)
model.summary()

# Compile & run training
model.compile(optimizer='rmsprop', loss='categorical_crossentropy')
# Note that `decoder_target_data` needs to be one-hot encoded,
# rather than sequences of integers like `decoder_input_data`!
model.fit([encoder_input_data, decoder_input_data],
          decoder_target_data,
          batch_size=batch_size,
          epochs=epochs,
          shuffle=True,
          validation_split=0.05)

我的模型总结如下：

Model: "functional_1"
__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
input_1 (InputLayer)            [(None, 13)]         0                                            
__________________________________________________________________________________________________
input_2 (InputLayer)            [(None, 15)]         0                                            
__________________________________________________________________________________________________
embedding (Embedding)           (None, 13, 512)      65536       input_1[0][0]                    
__________________________________________________________________________________________________
embedding_1 (Embedding)         (None, 15, 512)      65536       input_2[0][0]                    
__________________________________________________________________________________________________
lstm (LSTM)                     [(None, 512), (None, 2099200     embedding[0][0]                  
__________________________________________________________________________________________________
lstm_1 (LSTM)                   (None, 15, 512)      2099200     embedding_1[0][0]                
                                                                 lstm[0][1]                       
                                                                 lstm[0][2]                       
__________________________________________________________________________________________________
dense (Dense)                   (None, 15, 128)      65664       lstm_1[0][0]                     
==================================================================================================
Total params: 4,395,136
Trainable params: 4,395,136
Non-trainable params: 0
__________________________________________________________________________________________________

• • 编辑**

我正在按以下方式格式化数据：

for i, text, in enumerate(input_texts):
    words = text.split() #text is a sentence 
    for t, word in enumerate(words):
        encoder_input_data[i, t, input_dict[word]] = 1.

对于这样的命令decoder_input_data[:2]，给出：

array([[[0., 1., 0., ..., 0., 0., 0.],
        [0., 0., 1., ..., 0., 0., 0.],
        [0., 0., 0., ..., 0., 0., 0.],
        ...,
        [0., 0., 0., ..., 0., 0., 0.],
        [0., 0., 0., ..., 0., 0., 0.],
        [0., 0., 0., ..., 0., 0., 0.]],
       [[0., 0., 0., ..., 0., 0., 0.],
        [0., 0., 1., ..., 0., 0., 0.],
        [0., 0., 0., ..., 0., 0., 0.],
        ...,
        [0., 0., 0., ..., 0., 0., 0.],
        [0., 0., 0., ..., 0., 0., 0.],
        [0., 0., 0., ..., 0., 0., 0.]]], dtype=float32)