FA-99176 / Checkpoint tensor mapping / Open access

Checkpoint mapping: optional buffers are treated as required parameters · case 01

optional buffers are treated as required parameters.

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

ROOT CAUSE

Persistent buffers are considered required model weights.

THE FAILURE

Persistent buffers are considered required model weights.

Unsuccessful approach: Ignoring optional names only on the expected side still flags supplied buffers as unexpected.

Case contract

Report missing required names and unexpected incoming names; optional expected names may be absent.

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(required, incoming, optional):
    return (sorted(set(required)-set(incoming)), sorted(set(incoming)-set(required)))
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('missing optional', solve(*(['w'], ['w'], ['b'])), ([], []))
check('missing required', solve(*(['w', 'b'], ['b'], [])), (['w'], []))
check('optional incoming', solve(*(['w'], ['w', 'b'], ['b'])), ([], []))
check('extra', solve(*(['w'], ['w', 'x'], [])), ([], ['x']))
check('empty', solve(*([], [], [])), ([], []))
check('optional only absent', solve(*([], [], ['b'])), ([], []))
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 fixture	Actual	Expected	Outcome
missing optional	`[[], []]`	`[[], []]`	Passed
missing required	`[['w'], []]`	`[['w'], []]`	Passed
optional incoming	`[[], ['b']]`	`[[], []]`	Failed
extra	`[[], ['x']]`	`[[], ['x']]`	Passed
empty	`[[], []]`	`[[], []]`	Passed
optional only absent	`[[], []]`	`[[], []]`	Passed

SHA-256 / e5e42307713604aa534d744d70cc3dc1cd6094f71356c0f7f59b5f094cd413ec

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(required, incoming, optional):
    return (sorted(set(required)-set(incoming)-set(optional)), sorted(set(incoming)-set(required)))
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('missing optional', solve(*(['w'], ['w'], ['b'])), ([], []))
check('missing required', solve(*(['w', 'b'], ['b'], [])), (['w'], []))
check('optional incoming', solve(*(['w'], ['w', 'b'], ['b'])), ([], []))
check('extra', solve(*(['w'], ['w', 'x'], [])), ([], ['x']))
check('empty', solve(*([], [], [])), ([], []))
check('optional only absent', solve(*([], [], ['b'])), ([], []))
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 fixture	Actual	Expected	Outcome
missing optional	`[[], []]`	`[[], []]`	Passed
missing required	`[['w'], []]`	`[['w'], []]`	Passed
optional incoming	`[[], ['b']]`	`[[], []]`	Failed
extra	`[[], ['x']]`	`[[], ['x']]`	Passed
empty	`[[], []]`	`[[], []]`	Passed
optional only absent	`[[], []]`	`[[], []]`	Passed

SHA-256 / 3a26d504c7938666bf28397425e9c3144b915679bbfc8e17438024416b2be0b1

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.

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 2026-09-30T16:48:21.242536+00:00.

Case digest / 09af55feac616eb0559d01aff5875215c98a1e61418c9519b4fb77a0e6b4bedf