Implement the RMSProp optimizer to train a linear regression model using full-batch gradient descent with MSE loss.
