addmv_() allows resizing the tensor it operates on and produces wrong results #55589
Description
a = torch.randn((1,))
b = torch.randn((3, 4))
c = torch.randn((4,))
# this should error out
a.addmv_(b, c)
: tensor([ 2.2854, 1.4802, -1.4424])
This operation should fail. Discovered by @Lilyjjo (see #55539).
cc @ezyang @gchanan @zou3519 @bdhirsh @jbschlosser @anjali411 @jianyuh @nikitaved @pearu @mruberry @heitorschueroff @walterddr @IvanYashchuk