I have created a LSTM model for keyword prediction. I am using RMSE optimizer for training. I observe that the train and test accuracy decreases at first and then fluctuates without much difference in avg accuracy. I am trying to understand two things here-
- Why is the accuracy ( test and train ) decreasing at the start (epochs 0 -200). Also why isn't there any increasing trend in accuracy with number of epochs ( even with training dataset )
Why is there huge fluctuation with every epoch?
Any help will be appreciated.