tfmot.quantization.keras.experimental.default_n_bit.DefaultNBitQuantizeRegistry

QuantizationRegistry for built-in Keras classes for default 8-bit scheme.

Inherits From: QuantizeRegistry

tfmot.quantization.keras.experimental.default_n_bit.DefaultNBitQuantizeRegistry(
    disable_per_axis=False, num_bits_weight=8, num_bits_activation=8
)

Methods

get_quantize_config

View source

get_quantize_config(
    layer
)

Returns the quantization config for the given layer.

Args
layer input layer to return quantize config for.

Returns
Returns the QuantizeConfig for the given layer.

supports

View source

supports(
    layer
)

Returns whether the registry supports this layer type.

Args
layer The layer to check for support.

Returns
True/False whether the layer type is supported.