AttributeError: module 'keras.backend.tensorflow_backend' has no attribute '_LEARNING_PHASE' about recurrentshop HOT 4 CLOSED

Same here. The problem arises when using Dropout or possibly any other layer that has different training and testing time behavior. I tried to fix it myself, but ended up getting even deeper in trouble.

from recurrentshop.

I got this error then use something like:

model = Sequential()
model.add(Seq2Seq(input_shape=(inpLen, inpDim), output_length=inpLen, output_dim=inpDim, hidden_dim = hiddenDim, dropout=dropOut ,depth=2))
model.add(Activation('softmax'))
model.compile(loss='categorical_crossentropy', optimizer='adam')

from recurrentshop.

Any ideas @farizrahman4u ?
I got this problem when assigning the ground_truth value in the recurrent container's call.

from recurrentshop.

All fixed. See docs.

from recurrentshop.