quantize_model

Applicability

Product	Supported
Atlas A3 training series products/Atlas A3 inference series products	√
Atlas A2 training products/Atlas A2 inference products	√
Atlas 200I/500 A2 inference product	√
Atlas inference series products	√
Atlas training products	√

Description

Quantizes a graph based on the quantization configuration file, inserts the weight and activation quantization operators, generates a quantization factor record file record_file, and returns an ONNX model ready for calibration.

Prototype

quantize_model(config_file, model_file, modified_onnx_file, record_file)

Parameters

Parameter	Input/Output	Description
config_file	Input	User-generated distillation configuration file, which is used to specify the configuration of the quantization layer in the model network. A string.
model_file	Input	Original ONNX model file or updated model generated by the create_quant_config API. A string.
modified_onnx_file	Input	Name of the file for storing the ONNX calibration model for activation quantization. A string.
record_file	Input	Path (including the file name) of the quantization factor record file. A string.

Returns

None

Example

import amct_onnx as amct

model_file = "resnet101.onnx"
scale_offset_record_file = os.path.join(TMP, 'scale_offset_record.txt')
modified_model = os.path.join(TMP, 'modified_model.onnx')
config_file="./configs/config.json"
# Insert the quantization API.
amct.quantize_model(config_file,
                    model_file,
                    modified_model,
                    scale_offset_record_file)

Parent topic: PTQ APIs