Rate this Page
★★★★★

DeQuantStub#

classtorch.ao.quantization.DeQuantStub(qconfig=None)[source]#

Dequantize stub module, before calibration, this is same as identity,this will be swapped asnnq.DeQuantize inconvert.

Parameters

qconfig (Optional[Any]) – quantization configuration for the tensor,if qconfig is not provided, we will get qconfig from parent modules