Modify convolution operators to improve bias handling

Accumulator size moves to an enumerated attribute, out_t for floating-point changes
to be the size of the input. Bias for floating-point also becomes the bit width
of the input type.

Signed-off-by: Eric Kunze <eric.kunze@arm.com>
Change-Id: I7369417adbb1106ce34a1978e7f511a30272c318