class ov::pass::ConvToBinaryConv¶
Overview¶
This transformation converts Convolution to BinaryConvolution under following conditions: More…
#include <conv_to_binary_conv.hpp>
class ConvToBinaryConv: public ov::pass::MatcherPass
{
public:
// methods
OPENVINO_RTTI("ConvToBinaryConv", "0");
};Inherited Members¶
public:
// typedefs
typedef DiscreteTypeInfo type_info_t;
// methods
bool get_property(const PassPropertyMask& prop_mask) const;
void set_name(const std::string& name);
std::string get_name() const;
void set_callback(const param_callback& callback);
virtual void set_pass_config(const std::shared_ptr<PassConfig>& pass_config);
std::shared_ptr<PassConfig> get_pass_config();
bool m_transformation_callback(const std::shared_ptr<const Node>& node);
bool transformation_callback(const std::shared_ptr<const Node>& node);
virtual const type_info_t& get_type_info() const = 0;
OPENVINO_RTTI("ov::pass::MatcherPass");
MatcherPass& operator = (const MatcherPass&);
bool apply(std::shared_ptr<ov::Node> node);
template <typename T, class... Args>
std::shared_ptr<T> register_new_node(Args&&... args);
template <typename T>
std::shared_ptr<T> register_new_node(const std::shared_ptr<T>& node);
std::shared_ptr<ov::Node> register_new_node_(const std::shared_ptr<ov::Node>& node);
const std::vector<std::shared_ptr<ov::Node>>& get_new_nodes();
void clear_new_nodes();
std::shared_ptr<pattern::Matcher> get_matcher();Detailed Documentation¶
This transformation converts Convolution to BinaryConvolution under following conditions:
first input to Convolution is a FakeQuantize with levels==2 with output low,high being either (0, 1) or (-1, 1)
second input (weights) is a constant with values -1 or 1 The transformation also converts weights to binary Constant (with ‘u1’ type) For example, when output_low is equal to 0 and output_high is equal to 1, following graph
.... .... out_low out_high | | | | +--------------------------+ +-------------------------------------+ | FakeQuantize (levels==2) | | Constant | | (on activations) | | (weights containing -1 or 1 values) | +--------------------------+ +-------------------------------------+ | | | | ----------------- ------------------- | | v v +-------------+ | Convolution | +-------------+ | v
is transformed to:
.... .... out_low out_high | | | | +--------------------------+ +---------------------------------+ | FakeQuantize (levels==2) | | Constant (with u1 type) | | (on activations) | | (with u1 type - binary weights) | +--------------------------+ +---------------------------------+ | | | | ----------------- ------------------- | | v v +-------------------+ | BinaryConvolution | +-------------------+ | v +------------+ +----------------------------------------------------+ | | | Constant | | Add | <---| (weights from original graph, | | | | sum-reduced over [1,..., len(weights.shape)] axes | +------------+ +----------------------------------------------------+ | v +------------+ +-----+ | Multiply | <---| 0.5 | +------------+ +-----+ | v