Speeding Up a Slow Deep Learning Model Without Raising Cost

Question 1

A deep learning model is created with multiple parameters and layers. It is performing well but the predictions are very slow. The model needs to be modified so that the performance does not deteriorate significantly but the predictions become faster without increasing the cost. Which of the given options should be used in this case? Note: Assume that the accuracy of the model is not affected.