Generating Operator Model Files in Batches

python3 run_generate_model.py -m <mode> -t <npu_type> -p <pipeline> -pool <pool_size>

<mode>: algorithm mode. The value can be All, PCAR, Flat, IVFPQ, SQ8, IVFSQ8, INT8, or IVFINT8. Use commas (,) to separate multiple values, for example, python3 run_generate_model.py -m PCAR,IVFSQ8. By default, all modes are selected. You can directly run the python3 run_generate_model.py command.

<npu_type>: hardware type. Currently, <npu_type> supports the Ascend 310 AI Processor and Atlas inference products (Atlas 300I Pro inference card). The default value is 310.

<pipeline>: whether to use multi-thread parallel pipeline to generate an operator model. The default value is true.

<pool_size>: size of the process pool for multi-process scheduling of operators generated in batches.

• Run the command to obtain multiple groups of operator model files. Before running the command, you need to modify the para_table.xml file in the current directory and fill the required parameters in the table.
• 0 ≤ pool_size ≤ 32