Is it feasible to do the prediction without running the model everytime, just by calling the equation of my train model to predict the test dataset?

Question

I am running a linear equation and random forest model and every time i have to run a huge train data set to generate the model and eventually using the model to predict the