FlatAT

python3 flat_at_generate_model.py --cores <core_num> -d <dim> -c <code_num> -p <process_id> -t <npu_type>

<dim>: input feature vector dimension. The default value is 64.

<code_num>: number of base library features to be compared with the input features. The default value is 8192.

<core_num>: number of AI Cores of the Ascend AI Processor. The default value is 2. If this parameter is not specified, set this parameter based on <npu_type>. When npu_type is set to 310, set <core_num> to 2. When npu_type is set to 310P, set <core_num> to 8.

<process_id>: ID of the process for multi-process scheduling of operators generated in batches. The default value is 0, and you do not need to set this parameter.

<npu_type>: hardware form. Currently, Atlas 200/300/500 inference product and Atlas inference product are supported. The value can be 310 (default) or 310P.

--help | -h: help information.

Run the command to obtain a group of operator model files.

The FlatAT operator is used in the IVF scenario to reduce the time required for train and add functions.

• dim ∈ {64, 128, 256}
• code_num ∈ {1024, 2048, 4096, 8192, 16384, 32768}