Keras LSTM 转换为不同形状输入层中的 Core ML 结果

Question

我已经转换一个Keras模型到使用核心ML模型coremltools。原始的 Keras 模型具有以下架构：

__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
word_embeddings (InputLayer)    (None, 30)           0                                            
__________________________________________________________________________________________________
embedding (Embedding)           (None, 30, 256)      12800000    input_1[0][0]                    
__________________________________________________________________________________________________
activation_1 (Activation)       (None, 30, 256)      0           embedding[0][0]                  
__________________________________________________________________________________________________
bi_lstm_0 (Bidirectional)       (None, 30, 1024)     3149824     activation_1[0][0]               
__________________________________________________________________________________________________
bi_lstm_1 (Bidirectional)       (None, 30, 1024)     6295552     bi_lstm_0[0][0]                  
__________________________________________________________________________________________________
concatenate_1 (Concatenate)     (None, 30, 2304)     0           bi_lstm_1[0][0]                  
                                                                 bi_lstm_0[0][0]                  
                                                                 activation_1[0][0]               
__________________________________________________________________________________________________
attlayer (AttentionWeightedAver (None, 2304)         2304        concatenate_1[0][0]              
__________________________________________________________________________________________________
softmax (Dense)                 (None, 64)           147520      attlayer[0][0]                   
==================================================================================================

我可以使用以下输入对 Keras/Python 中的模型进行推理：

__________________________________________________________________________________________________
Layer (type)                    Output Shape         Param #     Connected to                     
==================================================================================================
word_embeddings (InputLayer)    (None, 30)           0                                            
__________________________________________________________________________________________________
embedding (Embedding)           (None, 30, 256)      12800000    input_1[0][0]                    
__________________________________________________________________________________________________
activation_1 (Activation)       (None, 30, 256)      0           embedding[0][0]                  
__________________________________________________________________________________________________
bi_lstm_0 (Bidirectional)       (None, 30, 1024)     3149824     activation_1[0][0]               
__________________________________________________________________________________________________
bi_lstm_1 (Bidirectional)       (None, 30, 1024)     6295552     bi_lstm_0[0][0]                  
__________________________________________________________________________________________________
concatenate_1 (Concatenate)     (None, 30, 2304)     0           bi_lstm_1[0][0]                  
                                                                 bi_lstm_0[0][0]                  
                                                                 activation_1[0][0]               
__________________________________________________________________________________________________
attlayer (AttentionWeightedAver (None, 2304)         2304        concatenate_1[0][0]              
__________________________________________________________________________________________________
softmax (Dense)                 (None, 64)           147520      attlayer[0][0]                   
==================================================================================================

但是，转换后，我的 Core ML 模型显示输入层的以下形状。

input {
  name: "word_embeddings"
  type {
    multiArrayType {
      shape: 1
      dataType: FLOAT32
    }
  }
}

为什么输入层 ( word_embeddings) 在 Core ML 转换后失去了形状？我希望它的类型显示MultiArray (Float32 1 x 30). 理想情况下，我希望能够像以前一样传入一个完整的 30 元素向量。

我已经阅读了Apple 的 Core ML LSTM 文档，它表明我可能需要一次重复调用model.prediction(..)一个元素，捕获每个预测的输出状态并将它们作为输入传递给下一个，直到我到达结束完整序列（所有 30 个元素）。或者，我可以利用Core ML Batch API使这更容易吗？

这是必要的吗？

对于上下文，该模型采用“句子”，即 30 个单词的列表（标记为唯一整数）并输出“表情符号”分类，即 64 个可能值之一。

使用 coremltools == 3.3、keras == 2.3.1 和 Python 2.7.17 在 Colab 中运行转换。