reverse the scale of the test outcome in the LSTM

Question

I am using standardized predictors in training set to train an LSTM model. After I predict the outcome in test set, I need to reverse the predicted score back to the origina