quantize_preprocess

Applicability

Product	Supported
Atlas 350 Accelerator Card	√
Atlas A3 training product/Atlas A3 inference product	√
Atlas A2 training product/Atlas A2 inference product	√
Atlas 200I/500 A2 inference product	√
Atlas inference product	√
Atlas training product	√

Description

Preprocesses the quantization of a graph based on the quantization configuration file (currently, only balanced quantization is supported), inserts the balanced quantization operators, and generates a balance factor record file record_file that is to be read in the subsequent quantize_model phase.

Prototype

quantized_preprocess(graph, config_file, record_file, outputs=None)

Parameters

Parameter	Input/Output	Description
graph	Input	tf.Graph of the model for quantization. A tf.Graph.
config_file	Input	User-generated quantization configuration file, which specifies the configuration of each layer to be quantized in the tf.Graph. A string.
record_file	Input	Path (including the file name) of the balance factor record file. A string.
outputs	Input	List of output operators in a graph. When the output nodes change due to graph modification, this list is updated accordingly. A list. Default: None

Example

import amct_tensorflow as amct
# Build a network to be quantized.
network = build_network()

# Insert the quantization API.
amct.quantized_preprocess(
      graph=tf.get_default_graph(),
      config_file="./configs/config.json",
      record_file="./record_scale_offset.txt")

Parent topic: PTQ APIs