Hello every one,

actually, we should pay attention to the type of LSTM model we choose, in my case as it was a time series classification problem, I'd choose a stateful RNN.

In addition, the problem I had was with the dimension of the test, train samples, so in this case, we should pay attention to the dimension of our arrays [use the reshape np function (x_train: x, y, z; y_ train: x,y; x_test: x,y,z & y_test x,y)] and be sure to pass them as np.arrays.

Finally, we have to choose the right loss function, in my case the binary_crossentropy function and be sure that the number of passed samples is multiple of the batch size we pass to the function fit (tweak the train_test_split 3rd function parameter).

Voilà, good luck for everyone.

https://stackoverflow.com/questions/45839016/lstm-keras-package-incoherent-output

https://machinelearningmastery.com/text-generation-lstm-recurrent-neural-networks-python-keras/

Hello

You question is not clear. reform your question to get answers. Make sure that you read and follow the approriate tutorials so that you will not get repeated answers.

Cordially