Class performing element-wise linear quantization. More...

Inheritance diagram for ngraph::op::v0::FakeQuantize:

Collaboration diagram for ngraph::op::v0::FakeQuantize:

Public Member Functions
	FakeQuantize (const Output< Node > &data, const Output< Node > &input_low, const Output< Node > &input_high, const Output< Node > &output_low, const Output< Node > &output_high, std::size_t levels, const AutoBroadcastSpec &auto_broadcast=AutoBroadcastSpec(AutoBroadcastType::NUMPY))
	Constructs a FakeQuantize operation node. More...

bool	visit_attributes (AttributeVisitor &visitor) override

virtual OutputVector	decompose_op () const override

virtual void	validate_and_infer_types () override

virtual std::shared_ptr< Node >	clone_with_new_inputs (const OutputVector &new_args) const override

std::size_t	get_levels () const

void	set_levels (std::size_t levels)

const AutoBroadcastSpec &	get_auto_broadcast () const

void	set_auto_broadcast (const AutoBroadcastSpec &auto_broadcast)

Detailed Description

Class performing element-wise linear quantization.

Note: Input floating point values are quantized into a discrete set of floating point values.

operation:

round((data - input_low) / (input_high - input_low) * (levels-1)) / (levels-1) * (output_high - output_low) + output_low

Constructor & Destructor Documentation

ngraph::op::v0::FakeQuantize::FakeQuantize	(	const Output< Node > &	data,
		const Output< Node > &	input_low,
		const Output< Node > &	input_high,
		const Output< Node > &	output_low,
		const Output< Node > &	output_high,
		std::size_t	levels,
		const AutoBroadcastSpec &	auto_broadcast = `AutoBroadcastSpec(AutoBroadcastType::NUMPY)`
	)

Constructs a FakeQuantize operation node.

Parameters

[in]	data	The input data tensor.
[in]	input_low	The minimum limit for input values.
[in]	input_high	The maximum limit for input values.
[in]	output_low	The minimum quantized value.
[in]	output_high	The maximum quantized value.
[in]	levels	The number of quantization levels.
[in]	auto_broadcast	AutoBroadcast mode to be used for broadcasting limit values

The documentation for this class was generated from the following file: