FAILURE MAP
← Case archive

FA-99171 / Checkpoint tensor mapping / Open access

Checkpoint mapping: unexpected checkpoint tensors are accepted · case 01

unexpected checkpoint tensors are accepted.

Verified by executionVariant 1 · 6 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

Only missing names are reported, concealing unexpected payloads.

THE FAILURE

Only missing names are reported, concealing unexpected payloads.

Unsuccessful approach: Reporting only the first extra name misses simultaneous unexpected entries.

Case contract

Return missing and unexpected names as sorted lists in a pair; names present in both sets are matched.

Why this case matters

A bounded offline model of checkpoint tensor bookkeeping in a training workflow.

1 / The failure

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(expected, incoming):
    return (sorted(set(expected)-set(incoming)), [])
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('extra many', solve(*(['a'], ['a', 'z', 'y'])), ([], ['y', 'z']))
check('missing', solve(*(['a', 'b'], ['a'])), (['b'], []))
check('exact', solve(*(['a'], ['a'])), ([], []))
check('both empty', solve(*([], [])), ([], []))
check('all extra', solve(*([], ['z', 'y'])), ([], ['y', 'z']))
check('mixed', solve(*(['a', 'b'], ['a', 'x'])), (['b'], ['x']))
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
extra many[[], []][[], ['y', 'z']]Failed
missing[['b'], []][['b'], []]Passed
exact[[], []][[], []]Passed
both empty[[], []][[], []]Passed
all extra[[], []][[], ['y', 'z']]Failed
mixed[['b'], []][['b'], ['x']]Failed

SHA-256 / 19eb465224b86eea67bc1870056d4cfcdd9685313a6e7d2fc20849a53d759e0f

2 / The unsuccessful fix

Exit 1
"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(expected, incoming):
    return (sorted(set(expected)-set(incoming)), sorted(set(incoming)-set(expected))[:1])
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('extra many', solve(*(['a'], ['a', 'z', 'y'])), ([], ['y', 'z']))
check('missing', solve(*(['a', 'b'], ['a'])), (['b'], []))
check('exact', solve(*(['a'], ['a'])), ([], []))
check('both empty', solve(*([], [])), ([], []))
check('all extra', solve(*([], ['z', 'y'])), ([], ['y', 'z']))
check('mixed', solve(*(['a', 'b'], ['a', 'x'])), (['b'], ['x']))
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)
Boundary fixtureActualExpectedOutcome
extra many[[], ['y']][[], ['y', 'z']]Failed
missing[['b'], []][['b'], []]Passed
exact[[], []][[], []]Passed
both empty[[], []][[], []]Passed
all extra[[], ['y']][[], ['y', 'z']]Failed
mixed[['b'], ['x']][['b'], ['x']]Passed

SHA-256 / 596d0648dff70ae8d25ccadd6e5fa96963599549a75cff0e0369a1a20de0fd47

HELD IN THE MEMBER ARCHIVE

The verified repair and its recorded checks are member-only.

This mechanism has 6 recorded checks per implementation. The open-access tier publishes the failure and the unsuccessful fix; the repaired source that passes every check, and the observations that prove it, are available to members.

Every case sharing this mechanism uses the same contract and the same repair, so this one record is held back for all of them.

Member access is invitation-based. Sign in with your invited account to inspect the repair.

Sign in to the archive ↗

Verification & scope

Inputs are already decoded Python values; this toy contract makes no claim about a framework, file format, or production checkpoint. This reproducer isolates one failure mechanism. Results cover the supplied fixtures. Variants within a family share a test contract and should remain grouped when constructing evaluation splits. Related mechanisms with a shared evaluation_group must also remain together; these controlled models are not independent production incidents.

Observations recorded using Python 3.12.14 at 2026-09-30T16:48:21.244030+00:00.

Case digest / 4eddc04a08c4f0b12edfc62a6382e9d7d6660b62bc8fad8536c355bed872da0d