The central window would be RECTANGULAR but instead of being vertically orientated would be in a plane that tilts toward the locomotive (from bottom to top) as the 'flat' lower front of the locomotive angles backwards at about waist height in the cab (kinda like any car windscreen, (if it were flat))
However, it looks like the lower centre of the cab of the only part that is vertical and perpendicular (vertically and transversely) to the long axis of the locomotive. It looks like each side, right and left of the center front panel also angles towards the locomotive (from centrally to the side). The distance between the central and side window frames appears to remain constant between (say,) the top and bottom.
Hence, (while remaining parallel to the centre window edges, from the front and angling away from the central window on each side) the non-central windows are not parallelograms but QUADRANGULAR in shape. The tops and bottoms of these windows are parallel to each other and only the central side of them is parallel to the central window vertical edges.
Now, if theres a curve on these panels, everything I just said is wrong for the non-central windows. So hopefully whoever designed the kit also came to the same conclusions.
And since the corners are rounded all these shapes are approximate
While all of this might be evident on the prototype, I doubt you would appreciate it scaled down 76 times, so it's likely the window will appear as a rhomboidal parallelogram with rounded corners (side are unequal length), not a rhombus (equal)
Sorry, I may have driven you completely mad now