get_default_qat_qconfig_mapping

class torch.ao.quantization.qconfig_mapping.get_default_qat_qconfig_mapping(backend='x86', version=1) [source]

Return the default QConfigMapping for quantization aware training.

Parameters

• backend (*) – the quantization backend for the default qconfig mapping, should be one of [“x86” (default), “fbgemm”, “qnnpack”, “onednn”]
• version (*) – the version for the default qconfig mapping

Return type

QConfigMapping