Class ov::pass::ConvertQuantizeDequantize

class ConvertQuantizeDequantize : public ov::pass::MatcherPass

ConvertQuantizeDequantize transformation replaces following graph: FakeQuantize->Convert->Convert->Subtract->Multiply with a single FakeQuantize. Restrictions:

• quantized data type must be i8, u8, i16, or u16

• ’levels’ attribute to FakeQuantize must be equal to 256 or 65536

• (output_low, output_high) must match the quantized data type range

• ’zero_point’ and ‘scale’ must be broadcastable to FakeQuantize’s output

• supports mixed precision: quantizer and dequantizer can use different floating-point types (e.g., fp32/fp16)