Class ov::pass::FakeQuantizeConcatFusion

class FakeQuantizeConcatFusion : public ov::pass::MatcherPass

FakeQuantizeConcatFusion replaces identical FakeQuantize operations on all inputs of a Concat followed by an identical output FakeQuantize with a single FakeQuantize after Concat.

Before:

data_0 -> FQ(a,b,c,d,l) -\ \ data_1 -> FQ(a,b,c,d,l) -—> Concat -> FQ(a,b,c,d,l) -> next / data_2 -> FQ(a,b,c,d,l) -/

After:

data_0 —————-—\ \ data_1 ———————> Concat -> FQ(a,b,c,d,l) -> next / data_2 —————-—/