init

Applicability

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

Description

Initializes AMCT, saves the quantization factor record file, and parses the user model into a graph that can be used for the quantize_model and save_model API calls.

Prototype

graph = init(config_file, model_file, weights_file, scale_offset_record_file)

Parameters

Parameter	Input/Output	Description
config_file	Input	Quantization configuration file generated by the user. A string.
model_file	Input	Definition file (.prototxt) of the Caffe model. It must be the same as model_file in create_quant_config. A string. Restrictions: For layers for inference, the settings in LayerParameter in model_file must meet inference requirements. For example, use_global_stats of the BatchNorm layer must be set to 1.
weights_file	Input	Trained Caffe model file (.caffemodel). It must be the same as weights_file in create_quant_config. A string.
scale_offset_record_file	Input	Quantization factor record file. The existing file (if any) in the directory will be overwritten. A string.

Returns

Returns a graph parsed from the model.

Example

from amct_caffe import init
# Initialize the tool.
graph = init(config_file="./configs/config.json",
             model_file="./pretrained_model/model.prototxt",
             weights_file="./pretrained_model/model.caffemodel",
             scale_offset_record_file="./recording.txt")

Flush file: quantization factor record file. If quantization is performed again, the quantization factor record file output by the API will be overwritten.

Parent topic: PTQ APIs