如何在此LSTM示例代码中将可训练参数数量计算为335872？

Question

I got this sample code but can't figure out how to calculate the trainable parameters to be 335872? (showed in the following output)

I would appreciate it if anyone could help on this question. Thanks!

-------------------------code------------------------------------

input_shape = (None, num_encoder_tokens)

# Define an input sequence and process it.
  encoder_inputs = Input(shape=input_shape)
  encoder = LSTM(latent_dim, return_state=True)
  encoder_outputs, state_h, state_c = encoder(encoder_inputs)

# We discard `encoder_outputs` and only keep the states.
  encoder_states = [state_h, state_c]

  encoder_model = Model(encoder_inputs, encoder_states)
  encoder_model.summary(line_length=100)

  encoder_model.output_shape

---------------------output is as follows----------------------

---

Layer (type)           Output Shape                      Param #        
=================================================================================
input_2 (InputLayer)   (None, None, 71)                        0              
_________________________________________________________________________________
lstm_5 (LSTM)     [(None, 256), (None, 256), (None, 256)] 335872         
=================================================================================
Total params: 335,872
Trainable params: 335,872
Non-trainable params: 0
_________________________________________________________________________________
[(None, 256), (None, 256)]

Answer 1

I am assuming that you want to know how to train the model so that weight matrices , biases , etc. can be calculated.

The problem with your code is that you have only defined the architecture of your model. You haven't really compiled it. Do this in the end:

encoder_model.compile(loss='binary_crossentropy', optimizer='adam', metrics='binary_accuracy')

In the above line of code, loss , optimizer and metrics is up to you to choose depending on the type of your problem.