Created by: xiaolil1

We add this OP for re-quantization of INT8 tensor with different Scales. This OP will be used by INT8 concat OP, multi-inputs with different scale values. Tag: Not target for V1.3.