Documentation

Data Format Requirements

The msaccucmp.py script tool supports multiple comparison approaches. Keep the following naming conventions in mind when creating dump files and .npy files.

Table 1 Naming conventions of dump files

Data Type

Name Format

Note

Dump data file of the non-quantized offline model running on the Ascend AI Processor

{op_type}.{op_name}.{task_id}.{stream_id}.{timestamp}

The name formats are described as follows:

  • The names of op_type (operator type) and op_name (operator name) must comply with the A-Za-z0-9_- regular expression rules.
  • timestamp is a 16-bit timestamp.
  • Each of the following parameters consists of digits 0-9: task_id (task ID), stream_id (stream ID), output_index (Nth output)

Periods (.), forward slashes (/), backslashes (\), and spaces in op_type or op_name are replaced with underscores (_).

Dump data file of the quantized offline model running on the Ascend AI Processor

.npy file (Caffe, TensorFlow, or ONNX)

{op_name}.{output_index}.{timestamp}.npy
Parent topic: Appendixes