fix: calculation result in ex2

Hey rechtigi cooli sach, merci vell mol 😄 Gshet nachme rechti prokraschtinier projekt e de lernphase us ;)

Han grad mol drigluegt ond vergliche. Chas si, dass des 22 esch?

conv = torch.nn.Conv2d(
	in_channels=16,
	out_channels=32,
	kernel_size=(7, 7),
	stride=1,
	padding=0)

torchinfo.summary(conv, (8, 16, 28, 28))

get be mer:

==========================================================================================
Layer (type:depth-idx)                   Output Shape              Param #
==========================================================================================
Conv2d                                   [8, 32, 22, 22]           25,120
==========================================================================================
Total params: 25,120
Trainable params: 25,120
Non-trainable params: 0
Total mult-adds (Units.MEGABYTES): 97.26
==========================================================================================
Input size (MB): 0.40
Forward/backward pass size (MB): 0.99
Params size (MB): 0.10
Estimated Total Size (MB): 1.49
==========================================================================================

Ond wenni d'formle ned völlig verstand han müesst doch eigentli au 22 geh ned? 😄