Change the quantizer to match the behavior of the FBGEMM implementation (#20892)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/20892
FBGEMM uses 64 bit values. Need to change our implementation to match
Reviewed By: jerryzh168
Differential Revision: D15487664
fbshipit-source-id: 29cba26093c6f9aeafce14982c1ae12149e63562