torch.ao.quantization.stubs.DeQuantStub#

class torch.ao.quantization.stubs.DeQuantStub[源代码]#

Dequantize stub module, before calibration, this is same as identity, this will be swapped as nnq.DeQuantize in convert.