compare_results#
- classtorch.ao.quantization.compare_results(ref_results,actual_results)[source]#
Given two dict mapping fromdebug_handle_id (int) to list of tensorsreturn a map fromdebug_handle_id toNodeAccuracySummary that containscomparison information like SQNR, MSE etc.
- Parameters
ref_results (Dict[int,Tuple[str,object,List[torch.Tensor]]]) – reference results for each debug_handle_id
actual_results (Dict[int,Tuple[str,object,List[torch.Tensor]]]) – actual results for each debug_handle_id
- Returns
Dict[int, NodeAccuracySummary]
- Return type
dict[int, torch.ao.quantization.pt2e._numeric_debugger.NodeAccuracySummary]