pulsatrix::RMSNormModule Class Reference

y_{n,i} = gamma_i * x_{n,i} / rms(x_n), rms(x_n) = sqrt(mean_i(x_{n,i}^2) + eps), computed independently per batch row n. Batched ((N, num_features)), migrated from the original unbatched (rank-1) scope by campaign_exai_dl_library_batch_dimension_support – no mean-centering, no beta/bias term (RMSNorm's defining simplification vs. LayerNorm). More...

#include <rms_norm_module.hpp>

Inheritance diagram for pulsatrix::RMSNormModule:

Collaboration diagram for pulsatrix::RMSNormModule:

Public Member Functions
	RMSNormModule (int64_t num_features, DeviceBackend *backend, DeviceType device, float eps=1e-6f)
	Constructs an RMSNorm layer with zero-initialized gamma.
	RMSNormModule (int64_t num_features, DeviceBackend *backend)
	On backend's own device (backend->device()), default eps. Previously the device defaulted to Cpu regardless of backend (GPU-native-kernels Mission 0).
Tensor	backward (const Tensor &grad_output) override
	Computes the gradient w.r.t. this module's input, and accumulates gamma's gradient internally.
OpType	op_type () const override
	Normalization per charter's closed OpType set.
void	set_gamma (std::initializer_list< float > values)
	Overwrites the gamma buffer – test/initialization use only.
void	set_gamma (const std::vector< float > &values)
	Vector overload for runtime-sized sources – see Tensor's own vector ctor.
const Tensor &	gamma () const
const Tensor &	gamma_grad () const
Tensor	propagate_relevance (const Tensor &relevance_out, const LRPRuleConfig &config) override
	Identity-rule LRP relevance propagation (AttnLRP, Achtibat et al. 2024).
std::vector< NamedParamRef >	named_parameters () override
	This module's trainable parameters, each with its hierarchical name – the one place a module declares its parameters (roadmap FND-1).
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 bool	supports_lrp_rule (LRPRule rule) const
	Whether propagate_relevance() implements rule (no silent fallback: callers such as ExplainerContext::relevance_pass() throw rather than run a module on a rule it does not implement).
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. Called by forward() after precondition checks.

Detailed Description

y_{n,i} = gamma_i * x_{n,i} / rms(x_n), rms(x_n) = sqrt(mean_i(x_{n,i}^2) + eps), computed independently per batch row n. Batched ((N, num_features)), migrated from the original unbatched (rank-1) scope by campaign_exai_dl_library_batch_dimension_support – no mean-centering, no beta/bias term (RMSNorm's defining simplification vs. LayerNorm).

Note

propagate_relevance is an identity pass-through, cited to AttnLRP's normalization-layer treatment (Achtibat et al. 2024, already named in this project's charter for Attention/Transformer work) – the same pattern this codebase already uses for ReluModule/FlattenModule's parameterless/ elementwise-treated operations, not a new pattern invented for this module. backward() is the real, undetached training gradient – the identity-rule simplification applies only to relevance redistribution, never to the actual gradient used for parameter updates (mirrors LinearModule's own independent backward()/propagate_relevance() computations from the same cached forward state).

Constructor & Destructor Documentation

RMSNormModule() [1/2]

pulsatrix::RMSNormModule::RMSNormModule	(	int64_t	num_features,
		DeviceBackend *	backend,
		DeviceType	device,
		float	eps = 1e-6f
	)

Constructs an RMSNorm layer with zero-initialized gamma.

Parameters

num_features	Number of elements this layer normalizes over.
backend	Backend to allocate/compute through. Not owned; must outlive this module.
device	Which device every internal Tensor member is tagged as. Defaults to Cpu.
eps	Stabilizer added inside the sqrt, avoiding division by zero for an all-zero input. Defaults to 1e-6, matching LRPRuleConfig's own default epsilon order of magnitude.

Exceptions

std::invalid_argument

if num_features <= 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.

RMSNormModule() [2/2]

pulsatrix::RMSNormModule::RMSNormModule	(	int64_t	num_features,
		DeviceBackend *	backend
	)

On backend's own device (backend->device()), default eps. Previously the device defaulted to Cpu regardless of backend (GPU-native-kernels Mission 0).

Member Function Documentation

backward()

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

overridevirtual

Computes the gradient w.r.t. this module's input, and accumulates gamma's gradient internally.

Parameters

grad_output

Gradient w.r.t. this module's output. Must match num_features.

Returns

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

Exceptions

std::logic_error

if forward() has never been called.

Note

Device-generic: runs on Cpu, Cuda or Hip tensors (GPU-native-kernels Mission 2).

Implements pulsatrix::Module.

compute_device()

std::optional< DeviceType > pulsatrix::RMSNormModule::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::RMSNormModule::forward_impl ( const Tensor &  input )

overrideprotectedvirtual

The actual forward computation. Called by forward() after precondition checks.

Implements pulsatrix::Module.

gamma()

const Tensor & pulsatrix::RMSNormModule::gamma ( ) const

inline

gamma_grad()

const Tensor & pulsatrix::RMSNormModule::gamma_grad ( ) const

inline

named_parameters()

std::vector< NamedParamRef > pulsatrix::RMSNormModule::named_parameters ( )

inlineoverridevirtual

This module's trainable parameters, each with its hierarchical name – the one place a module declares its parameters (roadmap FND-1).

Returns

{name, {value, grad}} entries pointing directly at this module's own members, in a fixed order. Names are unique within the module tree. Default: empty (a parameterless module like ReluModule needs no override).

Note

Override this, not parameters(): saving, loading, freezing by name and optimizer parameter groups all key on these names.

Reimplemented from pulsatrix::Module.

op_type()

OpType pulsatrix::RMSNormModule::op_type ( ) const

inlineoverridevirtual

Normalization per charter's closed OpType set.

Implements pulsatrix::Module.

propagate_relevance()

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

overridevirtual

Identity-rule LRP relevance propagation (AttnLRP, Achtibat et al. 2024).

Parameters

relevance_out	Relevance at this module's output. Must match num_features.
config	Unused – the identity rule has no tunable parameter, unlike LinearModule's epsilon rule.

Returns

relevance_out, unchanged – conservation holds trivially by construction.

Exceptions

std::logic_error	if forward() has never been called (sequencing violation, consistent with every other module's LRP-entry-point precedent).
std::invalid_argument	if relevance_out's element count doesn't match num_features.

Implements pulsatrix::Module.

set_gamma() [1/2]

void pulsatrix::RMSNormModule::set_gamma

(

const std::vector< float > &

values

)

Vector overload for runtime-sized sources – see Tensor's own vector ctor.

set_gamma() [2/2]

void pulsatrix::RMSNormModule::set_gamma

(

std::initializer_list< float >

values

)

Overwrites the gamma buffer – test/initialization use only.

---

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

• include/pulsatrix/rms_norm_module.hpp