I am using an encoder-decoder architecture, with 2 layers each in the encoder and decoder and 128 neurons in each hidden layer. The inputs are in a 2D form: one column has the days and the other column has the time series dependent on the days (shape:(5780, 100, 2)). The output is a single value, that represents a particular day when the breaking point occurs (shape:(5780, 1, 1)). The breaking point is one of the time dependent values that is not of concern to me.
The problem is that after training, the output on all test data is almost exactly the same. I have experimented with high and low learning rates (ranging 1e-2 to 1e-5), the number of training epochs (300 to 3000).
What I haven't done is batch normalization or any kind of normalization, but I had done some operations with a similar model where the outputs are the a sequence of points in the second feature column, and it worked perfectly fine.
What could be going wrong here?
In addition, is there another model that is more suited to predicting this kind of an output with the test data that I have?