twodigit$/$train$\_$utils.py in this problom

In your projoct foldor, look as the part$2-$twodigit subtoldcr. Thcre you can find the filcs

mip.py and

conv.py$.$

Your main task hore is to complcse the code irside the method sain in these files

Do the folowing stcps:

Look at maLn method in coch file Idcontify the training and test data and labcle. How many images are inside

the train and test data? What is the sizc of cach imagc?

Look at the defirition of the MLP clacs in

mip.py$.$ Try to make scrac of what thoac lines are trying in

achicve. What is y$\_$train$[$a$]$ and y$\_$train$[1]?$

Look at train

utils.py$,$ particularly the run epoch function.

Now ghen the intuition you have bult wi:h the above stopa, complete the following tasks.

Fully connected network

$0$ points puasale fungrietual

This problom is nequincd but ungraded bccouse it taloes a lot of ncsources from the graders. Honcec plewse try

your best, subvit your code but do nat wovy if the grador docs not refum, and refor to the solutians after due

datc.

Complete the code main in

mip.py to buid a fuly$-$connocted model with a single hidden Izycr with E$4$ units.

For this, you nood to malos use of Linear loyora in PyTorch; Unifes in the provious single$-$digt case, hore you

nood in use two scparatc Incar output Isyers, in generate the two outputs $($corresponding to the first and

sccond digits$).$

Notc that we have prowided you with an implementation of rlatten $,$ which maps a higher dimensional terant

into an N xd onc, where N is the rumbcr of ssmplcs in your batch and $d$ is the length of the flattonod

dimension of your tensor is $Nhw,$ the flattencd dmenaion is $d-\{h*w\}|.$

Available Functions: Yau heve access to the tarch$\_$m modulc as mn $,$ to the tarch$\_$m$\_$tunctianal as $r$

and to the rlatten layce as rlatten ; No nocd to import arything.a inport torct

$4$ inpart torct $\backslash$an an ct

A part$\_$ra$\_$cata$\_$aL $+\text{'}../$a

$18$

$13$ batat$\_$siae $-$ b$=$

$12$ ro claccos $-2$e

$13$Prcss ESC then TAB or click outside of the code cditcr to exdt.

Incarract

Test resultsTraceback $\{$most recent cala Lant$\}$: tren part$2$ twadigat$\_$train utadx inport latckity data, r$-$n opoch, train madeL, flatter

NaduleMat TrurdZrrar: Na madule named 'part$2$ twaddRat'