Alex Padula PadLex

## torch_conv_layer_to_fully_connected.py
"""
The function `torch_conv_layer_to_affine` takes a `torch.nn.Conv2d` layer `conv`
and produces an equivalent `torch.nn.Linear` layer `fc`.

Specifically, this means that the following holds for `x` of a valid shape:
    torch.flatten(conv(x)) == fc(torch.flatten(x))

Or equivalently:
    conv(x) == fc(torch.flatten(x)).reshape(conv(x).shape)
	"""
	The function `torch_conv_layer_to_affine` takes a `torch.nn.Conv2d` layer `conv`
	and produces an equivalent `torch.nn.Linear` layer `fc`.

	Specifically, this means that the following holds for `x` of a valid shape:
	torch.flatten(conv(x)) == fc(torch.flatten(x))

	Or equivalently:
	conv(x) == fc(torch.flatten(x)).reshape(conv(x).shape)