📐gparse-test-inference-softmax

The groupant beside gparse/tests/inference_softmax.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

Components

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.

Parts

• gparse-test-inference-softmax-comparison

Model text

descriptant

Source-bound description

compare asserts outer/inner shape equality then counts differing sentence vectors and individual prediction tuples. Synthetic helper test changes XPOS only and catches shape panics. Ignored experiment loads external model/corpus, alternates 8 rounds, requires per-mode repeated tuples, compares hidden/arc abs deltas<0.01, checks scored equality.

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.

approximant

Forward approximant

compare asserts outer/inner shape equality then counts differing sentence vectors and individual prediction tuples. Synthetic helper test changes XPOS only and catches shape panics. Ignored experiment loads external model/corpus, alternates 8 rounds, requires per-mode repeated tuples, compares hidden/arc abs deltas<0.01, checks scored equality.

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

Comparison count zero means exact tuple equality only after shape validation, not equal probabilities. Existing timed external experiment is not executed.

Cost

O(number of tuple/string comparisons), plus cloned fixture; external experiment adds inference/loading/timing.

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

composition

Composition

compare asserts outer/inner shape equality then counts differing sentence vectors and individual prediction tuples. Synthetic helper test changes XPOS only and catches shape panics. Ignored experiment loads external model/corpus, alternates 8 rounds, requires per-mode repeated tuples, compares hidden/arc abs deltas<0.01, checks scored equality.

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

Comparison count zero means exact tuple equality only after shape validation, not equal probabilities. Existing timed external experiment is not executed.

A (sentences,tokens) delta admits all mismatch locations and types with those counts; zero delta admits any identical aligned tuple pair.

models

Typed model catalog

behavior_complement

Passing selected assertions excludes their stated counterexamples within the fixtures. External saved-model/corpus experiment remains unexecuted; success of helper checks cannot establish it.

reverse_reconstruction

A (sentences,tokens) delta admits all mismatch locations and types with those counts; zero delta admits any identical aligned tuple pair.

shape_boundary

Outer mismatch fails before pair traversal; inner mismatch before token comparison for that sentence.

effect_boundary

Selected test process owns local fixture tensors/state; explicit CUDA tests serialize. No teacher/API, dataset or external model reads selected.

assurance_scope

Compiled selected assertions establish only finite observations; ignored experiments and backend allocation failures remain gaps.

failure_precedence

Fixture creation and backend operations must succeed before assertions; a compile-only result is distinct from executed assertions.

tests

Checks and limits

compare_counts_tuple_changes_and_shape_errors

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 for exact names and unexecuted paths.

evidence

Execution evidence — 2026-10-31

Source: gparse/tests/inference_softmax.rs

• compare_counts_tuple_changes_and_shape_errors: twice debug, twice release, all four exits 0.

Controls

1 source-local compiled assertion mutants; 1 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

Only compare helper; original external model/corpus timing experiment not run.

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.

This page in Markdown → https://md.groupants.elamur.ai/free/en/mova/gparse/tests/inference_softmax.rs.groupant/