# Loss and accuracy remains constant in time series classification by LSTM

I have a time series data with a classification label of 1 and 0. I am using a LSTM model to classify the series by taking 100 consecutive timestamps as input with a single label.

Even after training for many epochs, the loss and accuracy remains constant without even a little variation. I have tried to change the optimizer(ADAM and SGD), learning rate and number of layers but still the issue remains.

Any idea on why this could be happening?