Overview

In ONNX scenarios, only accuracy comparison of non-quantized models is supported. The following table lists the data files required for comparison.

**Table 1** Data file requirements for comparison between a non-quantized original model and a non-quantized offline model
File	Description	How to Obtain
.npy file of the non-quantized original model	Benchmark data	Preparing ONNX .npy Files
.json file generated by converting the offline model file using ATC	Operator mapping obtaining	Preparing Network-wide Information Files
Dump data file of the non-quantized offline model running on the AI processor	Data to be compared	In the offline inference scenario, the methods for obtaining the dump data of the NPU environment are the same for different frameworks. For details, see the following: Preparing Dump Data Files of an Offline Model

Parent topic: Comparison Between GPU and NPU (ONNX Offline Inference)