tfmot.experimental.combine.Default8BitClusterPreserveQuantizeScheme

Default 8 bit Cluster Preserve Quantization Scheme.

Inherits From: Default8BitQuantizeScheme, QuantizeScheme

tfmot.experimental.combine.Default8BitClusterPreserveQuantizeScheme(
    preserve_sparsity=True
)

Used in the notebooks

Used in the guide
Cluster preserving quantization aware training (CQAT) Keras example Sparsity and cluster preserving quantization aware training (PCQAT) Keras example

Args
preserve_sparsity	the flag to enable prune-cluster-preserving QAT.

Methods

get_layout_transformer

View source

get_layout_transformer()

Returns the layout transforms for this scheme.

Returns
Returns the QuantizeLayoutTransform for this quantization scheme.

get_quantize_registry

View source

get_quantize_registry()

Returns the quantization registry for this scheme.

Returns
Returns the QuantizeRegistry for this quantization scheme.