Use the same $(*),{}_0$ and $C$ in $($d$).$ Be sure to use a new random permutation of the indices of the

inner loop for each iteration of the outer loop. Show the iteration$-$versus$-$accuracy $($outer iteration

and training accuracy$)$ curve in the same plot as that for batch gradient descent. The training

and test data$/$labels are provided in digits$\_$training$\_$data.csv$,$ digits$\_$training$\_$labels.csv$,$

digits$\_$test$\_$data.csv and digits$\_$test$\_$labels.csv$.$