BackendConfig

class torch.ao.quantization.backend_config.BackendConfig(name='')

Config that defines the set of patterns that can be quantized on a given backend, and how reference quantized models can be produced from these patterns.

A pattern in this context refers to a module, a functional, an operator, or a directed acyclic graph of the above. Each pattern supported on the target backend can be individually configured through BackendPatternConfig in terms of:

1. The supported input/output activation, weight, and bias data types

2. How observers and quant/dequant ops are inserted in order to construct the reference pattern, and

3. (Optionally) Fusion, QAT, and reference module mappings.

The format of the patterns is described in: https://github.com/pytorch/pytorch/blob/master/torch/ao/quantization/backend_config/README.md

Example usage:

import torch
from torch.ao.quantization.backend_config import BackendConfig, BackendPatternConfig, DTypeConfig, ObservationType
from torch.ao.quantization.fuser_method_mappings import reverse_sequential_wrapper2

weighted_int8_dtype_config = DTypeConfig(
    input_dtype=torch.quint8,
    output_dtype=torch.quint8,
    weight_dtype=torch.qint8,
    bias_type=torch.float)

linear_config = BackendPatternConfig(torch.nn.Linear)             .set_observation_type(ObservationType.OUTPUT_USE_DIFFERENT_OBSERVER_AS_INPUT)             .add_dtype_config(weighted_int8_dtype_config)             .set_root_module(torch.nn.Linear)             .set_qat_module(torch.nn.qat.Linear)             .set_reference_quantized_module(torch.nn.quantized._reference.Linear)

conv_relu_config = BackendPatternConfig((torch.nn.ReLU, torch.nn.Conv2d))             .set_observation_type(ObservationType.OUTPUT_USE_DIFFERENT_OBSERVER_AS_INPUT)             .add_dtype_config(weighted_int8_dtype_config)             .set_fused_module(torch.nn.intrinsic.ConvReLU2d)             .set_fuser_method(reverse_sequential_wrapper2(torch.nn.intrinsic.ConvReLU2d))

backend_config = BackendConfig("my_backend")             .set_backend_pattern_config(linear_config)             .set_backend_pattern_config(conv_relu_config)

classmethod from_dict(backend_config_dict)

Create a BackendConfig from a dictionary with the following items:

“name”: the name of the target backend

“configs”: a list of dictionaries that each represents a BackendPatternConfig

Return type:

BackendConfig

set_backend_pattern_config(config)

Set the config for an pattern that can be run on the target backend. This overrides any existing config for the given pattern.

Return type:

BackendConfig

set_backend_pattern_configs(configs)

Set the configs for patterns that can be run on the target backend. This overrides any existing config for a given pattern if it was previously registered already.

Return type:

BackendConfig

set_name(name)

Set the name of the target backend.

Return type:

BackendConfig

to_dict()

Convert this BackendConfig to a dictionary with the items described in from_dict().

Return type:

Dict[str, Any]