FakeQuantChannelWiseAbsMax¶

class paddle.nn.quant.quant_layers. FakeQuantChannelWiseAbsMax ( name: str | None = None, channel_num: int | None = None, quant_bits: int = 8, quant_axis: int = 0, dtype: DTypeLike = 'float32', quant_on_weight: bool = False, reduce_type: Literal[max] | None = None ) [source]

forward ( input: Tensor ) → Tensor

forward¶

Defines the computation performed at every call. Should be overridden by all subclasses.

Parameters

• *inputs (tuple) – unpacked tuple arguments

• **kwargs (dict) – unpacked dict arguments