Keras-输入层和嵌入层错误

Question

I am trying to make a model for Tv script generation and while running the following model, input layer and embedding layer error is occurring. I have tried running the model without these two lines and it works fine. Can someone please help me with the error?

embedding = 300
lstm_size = 128
vocab_size = len(vocab) #8420
seq_len = 100


model = Sequential()
model.add(Input((None, )))
model.add(Embedding(inp, input_dim = vocab_size, output_dim = embedding, 
input_length = 1000))
model.add(LSTM(lstm_size, return_sequences = True, return_state = True))
model.add(LSTM(lstm_size, return_sequences = True, return_state = True))
model.add(LSTM(lstm_size, return_sequences = True, return_state = True))
model.add(Flatten())
model.add(Dense(vocab_size))

---

TypeError                                 Traceback (most recent call last)
<ipython-input-66-695a9250515c> in <module>
 19 #model = Model(inp, out)
 20 model = Sequential()
---> 21 model.add(Input((None, )))
 22 model.add(Embedding(inp, input_dim = vocab_size, output_dim = embedding, input_length = 1000))
 23 model.add(LSTM(lstm_size, return_sequences = True, return_state = True))

~\Anaconda3\lib\site-packages\tensorflow\python\training\checkpointable\base.py in _method_wrapper(self, *args, **kwargs)
440     self._setattr_tracking = False  # pylint: disable=protected-access
441     try:
--> 442       method(self, *args, **kwargs)
443     finally:
444       self._setattr_tracking = previous_value  # pylint: disable=protected-access

~\Anaconda3\lib\site- packages\tensorflow\python\keras\engine\sequential.py in add(self, layer)
143       raise TypeError('The added layer must be '
144                       'an instance of class Layer. '
--> 145                       'Found: ' + str(layer))
146     self.built = False
147     set_inputs = False

TypeError: The added layer must be an instance of class Layer. Found: Tensor("input_37:0", shape=(?, ?), dtype=float32)


This is coming for the Input layer
and,

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-67-3c663f8df357> in <module>
 20 model = Sequential()
 21 #model.add(Input((None, )))
---> 22 model.add(Embedding(inp, input_dim = vocab_size, output_dim = embedding, input_length = 1000))
 23 model.add(LSTM(lstm_size, return_sequences = True, return_state = True))
 24 model.add(LSTM(lstm_size, return_sequences = True, return_state = True))

TypeError: __init__() got multiple values for argument 'input_dim'

this comes for embedding layer.

Answer 1

Input is not a Layer Object. Which is why you get the first error. You do not need to pass something like that with a call to Sequential(). Embedding() can be your first layer.

And the second error is because you are passing inp to it. The first value should either be inp or vocab_size but it cannot be both.

Basically,

embedding = 300
lstm_size = 128
vocab_size = len(vocab) #8420
seq_len = 100


model = Sequential()
model.add(Embedding(vocab_size, embedding, input_length = 1000))