---
title: "gparse-model — Mova code model"
description: "Mova code model of gparse/src/model.rs: components with their labels, obernants and the model text."
url: "https://groupants.elamur.ai/free/en/mova/gparse/src/model.rs.groupant/"
---
Elamur: [Mova](https://md.mova.elamur.ai/free/en/) · [Dictionary](https://md.dictionary.elamur.ai/free/en/) · [Research](https://md.research.elamur.ai/free/en/) · **Groupants**

This page as HTML: <https://groupants.elamur.ai/free/en/mova/gparse/src/model.rs.groupant/>

# gparse-model

The groupant beside gparse/src/model.rs in the Mova repository: what the code computes, what it never does, how to go back from its outputs, its invariants, cost and checks.

## Facts

- **kind:** groupant
- **status:** partial source bound model
- **source:** [`gparse/src/model.rs`](https://github.com/ElamurAI/mova-free/blob/main/gparse/src/model.rs)
- **source SHA-256:** `a380c81aa11433fd…`
- **repository folder:** [`gparse/src/model.rs.groupant`](https://github.com/ElamurAI/mova-free/tree/main/gparse/src/model.rs.groupant)
- **repository commit:** `48f2e6f74146`

> **Implementation material.** A Mova code model describes this code at one source revision: partial models, finite checks and their stated scope. It is not a reviewed general exposition, and its labels are shown as written, never raised.

## Components

| component | label | where |
|---|---|---|
| **approximant** | partial | [approximant](#approximant) |
| **ast** | partial | `syntax.json` |
| **composition** | partial | [composition](#composition) |
| **cost** | symbolic source derived | [invariant](#invariant) |
| **descriptant** | present | [descriptant](#descriptant) |
| **evidence** | executed finite checks | [evidence](#evidence) |
| **invariant** | partial | [invariant](#invariant) |
| **tests** | executed bounded checks | [tests](#tests) |

## Obernants

The typed relations around the forward model: the behaviour complement (what the code never does), the reverse reconstruction (every input compatible with an output) and specialised views.

| id | type | label | where |
|---|---|---|---|
| `behavior_complement` | behavior_complement | partial source derived model | [behavior_complement](#behavior-complement) |
| `reverse_reconstruction` | reverse_reconstruction | partial source derived model | [reverse_reconstruction](#reverse-reconstruction) |
| `gradient_transport` | gradient_transport | partial source derived model | [gradient_transport](#gradient-transport) |
| `configuration_precedence` | configuration_precedence | partial source derived model | [configuration_precedence](#configuration-precedence) |
| `tensor_coordinates` | tensor_coordinates | partial source derived model | [tensor_coordinates](#tensor-coordinates) |
| `finite_arithmetic` | finite_arithmetic | partial source derived model | [finite_arithmetic](#finite-arithmetic) |

## Parts

- [gparse-model-configuration](https://md.groupants.elamur.ai/free/en/mova/gparse/src/model.rs.groupant/configuration.groupant/)
- [gparse-model-attention](https://md.groupants.elamur.ai/free/en/mova/gparse/src/model.rs.groupant/attention.groupant/)
- [gparse-model-heads](https://md.groupants.elamur.ai/free/en/mova/gparse/src/model.rs.groupant/heads.groupant/)

## Model text


### descriptant

#### Source-bound description

Cfg loads defaults, valid numeric replacements and strict erf/tanh GELU; Net::new applies GP_GELU then initializes embedding/transformer/head parameters. forward_impl sums word/pre/suf/shape/position/prag embeddings; pre-LN blocks use softmax(QKᵀ/sqrt(dh)+keybias)V and residual feedforward. Arc scores dep_i W head_j+head_j u+keybias+role pull. rels gathers flat word/head states and sums bilinear and concatenated-linear logits; tags gather word states.

##### Epoch 2026-10-31

Prospective model was recorded before edits. Added tests use bounded synthetic fixtures; production algorithms unchanged. Coverage is partial, not proof. No trained corpus, live teacher or throughput experiment was executed.

##### Epoch 2026-11-04

Cfg has ngram (0 by default). save appends `ngram=N` only when N>0, so a reference configuration's text is unchanged; load parses it and fails on a malformed value. Net has an optional gram embedding (ngram × d, σ 0.02) built when ngram>0; forward adds the masked mean of the batch's n-gram embeddings (row 0 masked out, divided by max(count,1)) after the position embedding, and fails when a subword model gets a batch without n-grams.

### approximant

#### Forward approximant

Cfg loads defaults, valid numeric replacements and strict erf/tanh GELU; Net::new applies GP_GELU then initializes embedding/transformer/head parameters. forward_impl sums word/pre/suf/shape/position/prag embeddings; pre-LN blocks use softmax(QKᵀ/sqrt(dh)+keybias)V and residual feedforward. Arc scores dep_i W head_j+head_j u+keybias+role pull. rels gathers flat word/head states and sums bilinear and concatenated-linear logits; tags gather word states.

Domain: valid source-specific dimensions and resources stated in invariant; no automatic production substitution. Numerical observations are finite checks, not an end-to-end relative bound. Executable Rust surrogate: unsupported gap; existing exact tests are not surrogates. Future metric-specific caps may be 1/2/5/10%, never above 10%.

### invariant

#### Invariant and cost

LayerNorm uses mean, variance and epsilon 1e-5 through differentiable ops. Training fused route calls custom differentiable softmax; fused inference uses upstream no-backward operation. heads>0, d divisible heads, valid batch dimensions and embedding indices are assumed, not cfg-validated. Padding uses finite -1e9 bias, not mathematical -infinity.

##### Cost

Let B,T,d,L,H,f,a,r,N,X,U,K denote batch, padded length, encoder width, layers, heads, FF width, arc width, relation width, real rows and XPOS/UPOS/relation class counts. Forward encoder costs O(B*T*d + L*(B*T*d*d+B*T*T*d+B*T*d*f)); arc projection/scoring O(B*T*d*a+B*T*a*a+B*T*T*a); relation gather/projections/classifier O(N*d*r+N*r*r*K+N*r*K), tags O(N*d*(X+U)). Memory includes B*H*T*T attention cells, BTd encoder states and retained training graphs plus parameters/optimizer state. Embedding/final-LN BTd remains when L=0; two AFFIX*d tables remain parameter allocations. Dimension products and backend costs are not measured timing claims.

Allocation, finite arithmetic, platform dimensions and external failures are not universally certified.

### composition

#### Composition

Cfg loads defaults, valid numeric replacements and strict erf/tanh GELU; Net::new applies GP_GELU then initializes embedding/transformer/head parameters. forward_impl sums word/pre/suf/shape/position/prag embeddings; pre-LN blocks use softmax(QKᵀ/sqrt(dh)+keybias)V and residual feedforward. Arc scores dep_i W head_j+head_j u+keybias+role pull. rels gathers flat word/head states and sums bilinear and concatenated-linear logits; tags gather word states.

##### Scoped children

The byte-scoped children project distinct internal relations through the parent syntax file. The parent combines them in actual source order. Remaining declarations/tests outside child ranges belong to this parent. Their correctness does not establish external caller preconditions.

##### Interfaces and limits

LayerNorm uses mean, variance and epsilon 1e-5 through differentiable ops. Training fused route calls custom differentiable softmax; fused inference uses upstream no-backward operation. heads>0, d divisible heads, valid batch dimensions and embedding indices are assumed, not cfg-validated. Padding uses finite -1e9 bias, not mathematical -infinity.

Cfg::save identifies fields but input text is many-to-one due unknown keys, invalid numeric values and repeats. Network outputs admit parameter symmetries and many input feature collisions; invertible reconstruction is not claimed.

### models

## Typed model catalog

### behavior_complement

Training never selects the upstream no-backward softmax via public forward; train=false ignores GP_TRAIN_FUSED_SOFTMAX. Unknown GP_GELU fails before parameter allocation. Invalid numeric cfg strings preserve previous values rather than report error.

Since 2026-11-04: a model with ngram=0 has no gram table and ignores any n-grams in the batch; a subword model never runs silently without them (error); padding bucket 0 receives no gradient (checked).

### reverse_reconstruction

Cfg::save identifies fields but input text is many-to-one due unknown keys, invalid numeric values and repeats. Network outputs admit parameter symmetries and many input feature collisions; invertible reconstruction is not claimed.

### gradient_transport

Custom layer norm and training softmax retain Q/K/embedding gradients; inference fusion intentionally excludes Q/K gradient while value route survives.

### configuration_precedence

Defaults then ordered lines then GP_GELU override; invalid saved GELU aborts immediately even if a later line is valid.

### tensor_coordinates

QKV reshape B,T,3,heads,d/heads; transpose to B,heads,T,dh. Relations require flattened heads k*T+h.

### finite_arithmetic

Variance epsilon avoids zero denominator for finite constant rows; nonfinite weights, overflow and all-masked behavior have no universal finite guarantee.

### tests

#### Checks and limits

model::tests::embeddings_get_gradients; model::fused::which_ops_backprop; activation_configuration_roundtrips; CUDA gradient/GELU/synthetic optimizer checks; CPU prediction

Selected commands run twice debug and twice release; exact receipts and source-to-test mapping retained in wave30 gparse external run directory. Tests do not certify allocation failure, all NaN/overflow inputs, trained-model accuracy or corpus timing. Compiled mutants cover selected boundaries; per-source coverage is recorded in the external map, not presumed.

See [source execution evidence](https://md.groupants.elamur.ai/free/en/mova/gparse/src/model.rs.groupant/#evidence) for exact names and unexecuted paths.

Since 2026-11-04: model::subword::subword_input_learns_and_is_required (save/load round trip, unchanged reference text, malformed value error, gradient on the n-gram table but not on row 0, missing n-grams rejected).

### evidence

#### Execution evidence — 2026-10-31

Source: gparse/src/model.rs

- model::tests::embeddings_get_gradients: twice debug, twice release, all four exits 0.
- model::fused::which_ops_backprop: twice debug, twice release, all four exits 0.
- activation_configuration_roundtrips: twice debug, twice release, all four exits 0.
- cuda_values_and_derivatives: twice debug, twice release, all four exits 0.
- training_attention_keeps_gradients_and_inference_control_breaks_them: twice debug, twice release, all four exits 0.
- synthetic_cpu_prediction_preserves_slots_and_fallback: twice debug, twice release, all four exits 0.
- synthetic_single_optimizer_step: twice debug, twice release, all four exits 0.

##### Controls

1 source-local compiled assertion mutants; 3 observed control tests including dependency mutations. Mutant compile exit0, assertion exit101, byte-exact restoration and positive rerun exit0. Per-source controls are listed in external source-test-map.json, not implied for every source.

##### Limits

Finite synthetic CPU/CUDA observations; no trained-corpus quality or all-input numeric guarantee.

CUDA softmax comparisons use absolute tolerance2e-6. GELU grid maxima observed: values0.0004734993, derivatives0.00086838007, finite difference0.0000005534244635896357; respective strict caps5e-4/1e-3/1e-5. These are finite observations, not percentage profiles or throughput claims.

Receipts: external wave30 gparse run native-commands.json, results.json, controls.json and source-test-map.json. Sorted exact test names/outcomes match per profile; numeric diagnostics and timings need not be byte-identical. Strict scoped library Clippy reports ten pre-existing production lints; appended test blocks were formatted independently without production reformatting.