Consider the following equation for gradient descent:

$=$w$-$n$\text{'}$g$($L$,$ D$,$ w$)$

where g is the gradient function, L the loss function, and D$,$ the dataset and n denotes the learning rate and the corresponding pseudo$-$code for the mini$-$batch variant:

$1.$

$2.$

for i $=1$ to num$\_$iter $\{$

shuffle $($data$)$;

for batch in get$\_$batches $($data$,$ batch$\_$size$)\{$

grad$=$ eval$\_$gradient $($loss$\_$function, batch, w$)$;

W $=$ w$-$learning$\_$rate$*$ grad;

$\}$

Identify the line numbers in this code that can be parallelized. Argue how the said lines can be parallelized using a distributed shared memory model.