pulsatrix::MaxPool2DModule Class Reference

Max pooling, rank-4 (N, channels, H, W), matching Conv2DModule's convention. Stride fixed equal to kernel size (non-overlapping windows), no padding, no dilation – deferred until a real use case needs them, same minimal-cut discipline as Conv2DModule's original stride-1/no-padding scope cut. More...

#include <max_pool2d_module.hpp>

Inheritance diagram for pulsatrix::MaxPool2DModule:

Collaboration diagram for pulsatrix::MaxPool2DModule:

Public Member Functions
	MaxPool2DModule (int64_t kernel_h, int64_t kernel_w, DeviceBackend *backend)
	Constructs a max-pool layer.
Tensor	backward (const Tensor &grad_output) override
	Computes the gradient w.r.t. this module's input – only the cached argmax position within each window receives grad_output; every other position is 0.
OpType	op_type () const override
	Pooling per charter's closed OpType set.
Tensor	propagate_relevance (const Tensor &relevance_out, const LRPRuleConfig &config) override
	Winner-take-all LRP relevance propagation (Bach et al. 2015).
bool	supports_lrp_rule (LRPRule) const override
	Winner-take-all ignores the config: the same under every rule, so supports all of them.
std::optional< DeviceType >	compute_device () const override
	Where this layer computes, so forward() rejects an input on another device (FND-8).
Public Member Functions inherited from pulsatrix::Module
virtual	~Module ()=default
Tensor	forward (const Tensor &input)
	Runs this module's forward computation.
std::pair< Tensor, NodeId >	forward_traced (const Tensor &input, NodeId input_node, ComputationGraph &graph, Autograd &autograd)
	Runs forward() while also registering a ComputationGraph node (tagged with this module's op_type(), parented to input_node) and wiring an Autograd backward function that reuses this module's own backward() – the opt-in traced/explainable path, per Phase 2 Mission 0.
virtual std::vector< NamedParamRef >	named_parameters ()
	This module's trainable parameters, each with its hierarchical name – the one place a module declares its parameters (roadmap FND-1).
virtual std::vector< ParamRef >	parameters ()
	This module's trainable parameters and their gradients, for an optimizer to update uniformly across module types.
void	set_requires_grad (bool requires_grad, const std::string &prefix="")
	Freezes (false) or unfreezes (true) parameters by name (roadmap FND-2).
virtual void	set_training (bool training)
	Sets this module's training/eval mode. Defaults to training (matches every mainstream framework's Module default).
bool	is_training () const
	Whether this module is currently in training mode.

Protected Member Functions
Tensor	forward_impl (const Tensor &input) override
	The actual forward computation – per-window max, argmax cached per output element for backward()/propagate_relevance() to reuse.

Detailed Description

Max pooling, rank-4 (N, channels, H, W), matching Conv2DModule's convention. Stride fixed equal to kernel size (non-overlapping windows), no padding, no dilation – deferred until a real use case needs them, same minimal-cut discipline as Conv2DModule's original stride-1/no-padding scope cut.

Note

propagate_relevance is winner-take-all (Bach et al. 2015's supplementary treatment of max-pooling, the same rule iNNvestigate/zennit ship): all relevance at an output position flows to the single input position that was the argmax in forward(); every other position in that window gets zero. Conserves exactly by construction. backward() routes gradient the same way (only the argmax position receives grad_output; standard max-pool gradient semantics).

Constructor & Destructor Documentation

MaxPool2DModule()

pulsatrix::MaxPool2DModule::MaxPool2DModule	(	int64_t	kernel_h,
		int64_t	kernel_w,
		DeviceBackend *	backend
	)

Constructs a max-pool layer.

Parameters

kernel_h	Window height. Also the vertical stride (non-overlapping).
kernel_w	Window width. Also the horizontal stride (non-overlapping).
backend	Backend to allocate/compute through. Not owned; must outlive this module.

Exceptions

std::invalid_argument

if kernel_h <= 0 or kernel_w <= 0 – external boundary (construction arguments can originate from Phase 5's Python bindings with no upstream validation), per cpp_tdd/context_tdd_adversarial_boundary_testing.md.

Member Function Documentation

backward()

Tensor pulsatrix::MaxPool2DModule::backward ( const Tensor &  grad_output )

overridevirtual

Computes the gradient w.r.t. this module's input – only the cached argmax position within each window receives grad_output; every other position is 0.

Parameters

grad_output

Gradient w.r.t. this module's output. Must match the shape of the most recent forward() call's output.

Returns

Gradient w.r.t. this module's input.

Exceptions

std::logic_error	if forward() has never been called.
std::invalid_argument	if grad_output's shape doesn't match the cached forward output shape.

Note

Device-generic: runs on Cpu, Cuda or Hip tensors (GPU-native-kernels Mission 4).

Implements pulsatrix::Module.

compute_device()

std::optional< DeviceType > pulsatrix::MaxPool2DModule::compute_device ( ) const

inlineoverridevirtual

Where this layer computes, so forward() rejects an input on another device (FND-8).

Reimplemented from pulsatrix::Module.

forward_impl()

Tensor pulsatrix::MaxPool2DModule::forward_impl ( const Tensor &  input )

overrideprotectedvirtual

The actual forward computation – per-window max, argmax cached per output element for backward()/propagate_relevance() to reuse.

Exceptions

std::invalid_argument

if input isn't rank-4 (N, C, H, W), or the kernel is larger than the input (kernel_h > H or kernel_w > W).

Note

Device-generic: runs on Cpu, Cuda or Hip tensors (GPU-native-kernels Mission 4).

Implements pulsatrix::Module.

op_type()

OpType pulsatrix::MaxPool2DModule::op_type ( ) const

inlineoverridevirtual

Pooling per charter's closed OpType set.

Implements pulsatrix::Module.

propagate_relevance()

Tensor pulsatrix::MaxPool2DModule::propagate_relevance ( const Tensor &  relevance_out, const LRPRuleConfig &  config  )

overridevirtual

Winner-take-all LRP relevance propagation (Bach et al. 2015).

Parameters

relevance_out	Relevance at this module's output. Must match the shape of the most recent forward() call's output.
config	Unused – winner-take-all has no tunable parameter.

Returns

Relevance at this module's input: relevance_out's value at each window's cached argmax position, zero everywhere else. Conserves trivially by construction (no epsilon stabilizer needed).

Exceptions

std::logic_error	if forward() has never been called.
std::invalid_argument	if relevance_out's shape doesn't match the cached forward output shape.

Implements pulsatrix::Module.

supports_lrp_rule()

bool pulsatrix::MaxPool2DModule::supports_lrp_rule ( LRPRule  ) const

inlineoverridevirtual

Winner-take-all ignores the config: the same under every rule, so supports all of them.

Reimplemented from pulsatrix::Module.

---

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

• include/pulsatrix/max_pool2d_module.hpp