在 Keras LSTM 中添加双向会导致 RuntimeError（'您必须在使用之前编译模型。'）

Question

I'm doing:

self.model.add(Bidirectional(LSTM(lstm1_size, input_shape=(
        seq_length, feature_dim), return_sequences=True)))
self.model.add(Bidirectional(LSTM(lstm2_size, return_sequences=True)))
self.model.add(Bidirectional(LSTM(lstm3_size, return_sequences=True)))
self.model.add(Bidirectional(LSTM(lstm4_size, return_sequences=True)))
self.model.add(Dense(feature_dim, activation='relu'))

and I get the error:

RuntimeError('You must compile your model before using it.')

However, when I do the same thing without the Bidirectional , it works fine. What am I doing wrong?

Answer 1

it seems like a bug in your case because I tried the same its working normally.

import tensorflow as tf
from tensorflow.keras import Model
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from tensorflow.keras.layers import LSTM, Bidirectional

lstm1_size = 4
seq_length = 2
lstm2_size = 4
feature_dim = 1
lstm3_size = 4
lstm4_size = 4
model = Sequential()
model.add(Bidirectional(LSTM(lstm1_size, input_shape=(
        seq_length, feature_dim), return_sequences=True)))
model.add(Bidirectional(LSTM(lstm2_size, return_sequences=True)))
model.add(Bidirectional(LSTM(lstm3_size, return_sequences=True)))
model.add(Bidirectional(LSTM(lstm4_size, return_sequences=True)))
model.add(Dense(feature_dim, activation='relu'))

model.compile(loss='mean_squared_error', optimizer='adam')

please check if model is compiled within/outside class? if problem is solved already, please let us know as well.