Documentation

LinearQAT

Function Usage

Constructs the QAT operator of Linear.

Prototype

API for operator construction from scratch:

amct_pytorch.nn.module.quantization.linear.LinearQAT(in_features, out_features, bias, device, dtype, config)

API for construction based on the native operator:

amct_pytorch.nn.module.quantization.linear.LinearQAT.from_float(mod, config)

Command-Line Options

Table 1 Parameters in the API for operator construction from scratch

Option

Input/Output

Meaning

Restriction

in_features

Input

Number of input features.

An int.

This parameter is mandatory.

out_features

Input

Number of output features.

Type: int

This parameter is mandatory.

bias

Input

Indicates whether to enable bias items to participate in learning.

Type: bool

The default value is True.

device

Input

Running device.

Default: None

dtype

Input

Torch data type.

Torch data type. Only torch.float32 is supported.

config

Input

The following is a configuration example. For details about quantization configuration parameters, see Quantization Configuration Parameters.

config = {
    "retrain_enable":true,
    "retrain_data_config": {
        "dst_type": "INT8",
        "batch_num": 10,
        "fixed_min": False,
        "clip_min": -1.0,
        "clip_max": 1.0
    },
    "retrain_weight_config": {
        "dst_type": "INT8",
        "weights_retrain_algo": "arq_retrain",
        "channel_wise": False
    }
}

A dict.

Default: None
Table 2 Parameters in the API for construction based on native operators

Option

Input/Output

Description

Restriction

mod

Input

Native Linear operator to be quantized.

A torch.nn.Module.

config

Input

Quantization configuration.

The following is a configuration example. For details about quantization configuration parameters, see Quantization Configuration Parameters.

config = {
    "retrain_enable":true,
    "retrain_data_config": {
        "dst_type": "INT8",
        "batch_num": 10,
        "fixed_min": False,
        "clip_min": -1.0,
        "clip_max": 1.0
    },
    "retrain_weight_config": {
        "dst_type": "INT8",
        "weights_retrain_algo": "arq_retrain",
        "channel_wise": False
    }
}

A dict.

Default: None

Returns

A QAT operator of Linear for subsequent quantization perception training.

Examples

Construction from scratch:

1
2
3
4
from amct_pytorch.nn.module.quantization.linear import LinearQAT

LinearQAT(in_features=1, out_features=1, bias=True,
          device=None, dtype=None, config=None)

Construction based on the native operator:

1
2
3
4
5
6
import torch

from amct_pytorch.nn.module.quantization.linear import LinearQAT

linear_op = torch.nn.Linear(in_features=1, out_features=1, bias=True, device=None, dtype=None)
LinearQAT.from_float(mod=linear_op, config=None)
Parent topic: Single-operator mode