keras LSTM constructed input shape , called model input shape

0
votes

I used the LSTM MODEL of Keras.

input data shape : (None, 210, 4)

'210' is '21 samples * 10'

called input data shape : (None, 21, 4)

It was supposed to be a problem, but it wasn't. ; The predictions are well made.

Can I use this continuously?

How is this possible Why does this problem occur?

WARNING:tensorflow:Model was constructed with shape (None, 210, 4) for input KerasTensor(type_spec=TensorSpec(shape=(None, 210, 4), dtype=tf.float32, name='lstm_input'), name='lstm_input', description="created by layer 'lstm_input'"), but it was called on an input with incompatible shape (None, 21, 4).

1
kerasinputlstm

1 Answers

0
votes

Your question is not really clear, but in keras, the first dimension of your input is the number of instance in a batch. So you should change your code from

input_shape=(None,21,4)

to

input_shape=(10,21,4)

meaning you give 10 instance of a data which shape is (21,4)