Selects which LRP rule variant a Module::propagate_relevance() call uses. More...

#include <string>

Include dependency graph for lrp_rule_config.hpp:

This graph shows which files directly or indirectly include this file:

Classes
struct	pulsatrix::LRPRuleConfig
	Configuration for LRP relevance propagation: which rule a module applies and its hyperparameters. A module that does not implement the requested rule throws (see Module::supports_lrp_rule()) – the rule is never silently substituted. More...

Enumerations
enum class	pulsatrix::LRPRule { pulsatrix::Epsilon , pulsatrix::Gamma , pulsatrix::AlphaBeta , pulsatrix::ZBox }
	The LRP rule family a module applies. Semantics follow Zennit 1.0.0 exactly (Anders et al. 2021); every rule other than Epsilon is defined only for affine layers (LinearModule, Conv2DModule) – see Module::supports_lrp_rule(). More...

Functions
std::string	pulsatrix::lrp_rule_name (LRPRule rule)
	Lower-case rule name ("epsilon", "gamma", "alpha_beta", "zbox") for messages/metadata.

Detailed Description

Selects which LRP rule variant a Module::propagate_relevance() call uses.

See also: Module::propagate_relevance(), and the per-layer LRP rule implementations in Deep Learning Modules and Layers (e.g. LinearModule, Conv2DModule, TransformerBlock) – LRP rules live alongside each layer's forward/backward math, not in a separate file.

Classes