convert_qat_model

Function Usage

Converts a TensorFlow QAT model to a quantized model that serves for both deployment on Ascend AI Processor and accuracy simulation on the CPU and GPU.

Constraints

The original TensorFlow .pb model must have FakeQuantWithMinMaxVars and FakeQuantWithMinMaxVarsPerchannel operators.

Prototype

convert_qat_model(pb_model, outputs, save_path, record_file=None)

Command-Line Options

Parameter

Input/Return

Meaning

Restriction

pb_model

Input

Path of the QAT model to be adapted.

A string

outputs

Input

List of output operators of the graph.

A list.

record_file

Input

Path of the quantization factor record file (.txt) computed by the user.

A string

Default: None

save_path

Input

Model save path. Must include the prefix of the model name, for example, ./quantized_model/*model.

A string

Return Value

List of output operators of the graph.

Outputs

A .pb model file that serves for both accuracy simulation in the TensorFlow environment and offline inference on Ascend AI Processor.

When distillation is performed again, the preceding files output by this API will be overwritten.

Examples

1
2
import amct_tensorflow as amct
convert_qat_model(pb_model, outputs, save_path)