QuantConfig¶

class paddle.quantization. QuantConfig ( activation: QuanterFactory | None, weight: QuanterFactory | None ) [source]

Configure how to quantize a model or a part of the model. It will map each layer to an instance of SingleLayerConfig by the settings. It provides diverse methods to set the strategies of quantization.

Parameters

activation (QuanterFactory | None) – The global quantizer used to quantize the activations.
weight (QuanterFactory | None) – The global quantizer used to quantize the weights.

Examples

           >>> from paddle.quantization import QuantConfig
>>> from paddle.quantization.quanters import FakeQuanterWithAbsMaxObserver

>>> quanter = FakeQuanterWithAbsMaxObserver(moving_rate=0.9)
>>> q_config = QuantConfig(activation=quanter, weight=quanter)
>>> print(q_config)
Global config:
activation: FakeQuanterWithAbsMaxObserver(name=None,moving_rate=0.9,bit_length=8,dtype=float32)
weight: FakeQuanterWithAbsMaxObserver(name=None,moving_rate=0.9,bit_length=8,dtype=float32)

          

add_layer_config ( layer: Layer | list[Layer], activation: QuanterFactory | None = None, weight: QuanterFactory | None = None ) → None add_layer_config¶

Set the quantization config by layer. It has the highest priority among all the setting methods.

Parameters

layer (Layer|list[Layer]]) – One or a list of layers.
activation (QuanterFactory | None) – Quanter used for activations. Default is None.
weight (QuanterFactory | None) – Quanter used for weights. Default is None.

Examples

             >>> import paddle
>>> from paddle.nn import Linear
>>> from paddle.quantization import QuantConfig
>>> from paddle.quantization.quanters import FakeQuanterWithAbsMaxObserver

>>> class Model(paddle.nn.Layer):
...    def __init__(self):
...        super().__init__()
...        self.fc = Linear(576, 120)
>>> model = Model()
>>> quanter = FakeQuanterWithAbsMaxObserver(moving_rate=0.9)
>>> q_config = QuantConfig(activation=None, weight=None)
>>> q_config.add_layer_config([model.fc], activation=quanter, weight=quanter)
>>> 
>>> print(q_config)
Global config:
None
Layer prefix config:
{'linear_0': <paddle.quantization.config.SingleLayerConfig object at 0x7fe41a680ee0>}

            

add_name_config ( layer_name: str | list[str], activation: QuanterFactory | None = None, weight: QuanterFactory | None = None ) → None add_name_config¶

Set the quantization config by full name of layer. Its priority is lower than add_layer_config.

Parameters

layer_name (str|list[str]) – One or a list of layers’ full name.
activation (QuanterFactory | None) – Quanter used for activations. Default is None.
weight (QuanterFactory | None) – Quanter used for weights. Default is None.

Examples

             >>> import paddle
>>> from paddle.nn import Linear
>>> from paddle.quantization import QuantConfig
>>> from paddle.quantization.quanters import FakeQuanterWithAbsMaxObserver

>>> class Model(paddle.nn.Layer):
...     def __init__(self):
...         super().__init__()
...         self.fc = Linear(576, 120)
>>> model = Model()
>>> quanter = FakeQuanterWithAbsMaxObserver(moving_rate=0.9)
>>> q_config = QuantConfig(activation=None, weight=None)
>>> q_config.add_name_config([model.fc.full_name()], activation=quanter, weight=quanter)
>>> 
>>> print(q_config)
Global config:
None
Layer prefix config:
{'linear_0': <paddle.quantization.config.SingleLayerConfig object at 0x7fe41a680fd0>}

            

add_type_config ( layer_type: type[Layer] | list[type[Layer]], activation: QuanterFactory | None = None, weight: QuanterFactory | None = None ) → None add_type_config¶

Set the quantization config by the type of layer. The layer_type should be subclass of paddle.nn.Layer. Its priority is lower than add_layer_config and add_name_config.

Parameters

layer_type (type[Layer] | list[type[Layer]]) – One or a list of layers’ type. It should be subclass of
layer. (paddle.nn.Layer. Python build-in function type() can be used to get the type of a) –
activation (QuanterFactory | None) – Quanter used for activations. Default is None.
weight (QuanterFactory | None) – Quanter used for weights. Default is None.

Examples

             >>> import paddle
>>> from paddle.nn import Linear
>>> from paddle.quantization import QuantConfig
>>> from paddle.quantization.quanters import FakeQuanterWithAbsMaxObserver

>>> class Model(paddle.nn.Layer):
...     def __init__(self):
...         super().__init__()
...         self.fc = Linear(576, 120)
>>> model = Model()
>>> quanter = FakeQuanterWithAbsMaxObserver(moving_rate=0.9)
>>> q_config = QuantConfig(activation=None, weight=None)
>>> q_config.add_type_config([Linear], activation=quanter, weight=quanter)
>>> 
>>> print(q_config)
Global config:
None
Layer type config:
{<class 'paddle.nn.layer.common.Linear'>: <paddle.quantization.config.SingleLayerConfig object at 0x7fe41a680a60>}

            

add_qat_layer_mapping ( source: type[Layer], target: type[Layer] ) → None add_qat_layer_mapping¶

Add rules converting layers to simulated quantization layers before quantization-aware training. It will convert layers with type source to layers with type target. source and target should be subclass of paddle.nn.Layer. And a default mapping is provided by property default_qat_layer_mapping.

Parameters

source (type[Layer]) – The type of layers that will be converted.
target (type[Layer]) – The type of layers that will be converted to.

Examples

             >>> import paddle
>>> from paddle.nn import Conv2D
>>> from paddle.quantization import QuantConfig
>>> from paddle.quantization.quanters import FakeQuanterWithAbsMaxObserver
>>> quanter = FakeQuanterWithAbsMaxObserver(moving_rate=0.9)
>>> q_config = QuantConfig(activation=None, weight=None)
>>> class CustomizedQuantedConv2D(paddle.nn.Layer):
...     def forward(self, x):
...         pass
...         # add some code for quantization simulation
>>> q_config.add_qat_layer_mapping(Conv2D, CustomizedQuantedConv2D)

            

add_customized_leaf ( layer_type: type[Layer] ) → None add_customized_leaf¶

Declare the customized layer as leaf of model for quantization. The leaf layer is quantized as one layer. The sublayers of leaf layer will not be quantized.

Parameters: layer_type (type[Layer]) – The type of layer to be declared as leaf.

Examples

             >>> from paddle.nn import Sequential
>>> from paddle.quantization import QuantConfig
>>> from paddle.quantization.quanters import FakeQuanterWithAbsMaxObserver
>>> q_config = QuantConfig(activation=None, weight=None)
>>> q_config.add_customized_leaf(Sequential)

            

property customized_leaves : list[type[Layer]]: Get all the customized leaves.

details ( ) → str details¶: Get the formatted details of current config.

QuantConfig¶

add_layer_config¶

add_name_config¶

add_type_config¶

add_qat_layer_mapping¶

add_customized_leaf¶

details¶