Dump File Naming Conventions
The compare_vector.py accuracy comparison tool will be brought offline in later versions. You are advised to use msaccucmp.py in the current version.
The current version supports multiple comparison approaches. Keep the following naming conventions in mind when creating dump files and .npy files.
Data Type |
Naming Format |
|---|---|
Dump data of the non-quantized Caffe model |
{op_name}.{output_index}.{timestamp}.pb |
Dump data of the quantized Caffe model |
{op_name}.{output_index}.{timestamp}.quant |
Dump data of the non-quantized offline model running on the Ascend AI Processor |
{op_type}.{op_name}.{task_id}.{timestamp} |
Dump data of the quantized offline model running on the Ascend AI Processor |
{op_type}.{op_name}.{task_id}.{timestamp} |
Dump data of the non-quantized TensorFlow model |
{op_name}.{output_index}.{timestamp}.pb |
.npy file (Caffe or TensorFlow) |
{op_name}.{output_index}.{timestamp}.npy |
Where, op_type and op_name must comply with the A-Za-z0-9_- regular expression rule, timestamp is of 16 bits, and output_index and task_id are digits in the range 0–9.