FA-98051 / Sequence batch packing / Open access

Batch padding rounds down to a length multiple · case 01

Batch padding rounds down to a length multiple.

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

ROOT CAUSE

Floor division chooses the previous bucket when maximum length is not divisible.

THE FAILURE

Floor division chooses the previous bucket when maximum length is not divisible.

Unsuccessful approach: Adding a full extra bucket fixes nonmultiples but wastes a bucket for exact multiples.

Case contract

Return the smallest multiple of positive multiple that is at least the maximum length; empty input returns zero.

Why this case matters

This deterministic toy packing rule models a common boundary in preparing bounded training sequences.

1 / The failure

Exit 1

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

N = 1
observations = []
def solve(lengths, multiple):
    return (max(lengths)//multiple)*multiple if lengths else 0
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('between multiples', solve([2*N+1],N+1), 2*(N+1))
check('already multiple', solve([2*N],N), 2*N)
check('zero length', solve([0],N), 0)
check('empty', solve([],N), 0)
check('mixed maximum', solve([N,2*N+1,1],N), 3*N)
check('unit multiple', solve([N+1],1), N+1)
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
between multiples	`2`	`4`	Failed
already multiple	`2`	`2`	Passed
zero length	`0`	`0`	Passed
empty	`0`	`0`	Passed
mixed maximum	`3`	`3`	Passed
unit multiple	`2`	`2`	Passed

SHA-256 / 3fbdbec0066f6352f7a4cca62a83aab35de47b83f50aa76781f0adac6c072148

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(lengths, multiple):
    return ((max(lengths)+multiple-1)//multiple+1)*multiple if lengths else 0
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('between multiples', solve([2*N+1],N+1), 2*(N+1))
check('already multiple', solve([2*N],N), 2*N)
check('zero length', solve([0],N), 0)
check('empty', solve([],N), 0)
check('mixed maximum', solve([N,2*N+1,1],N), 3*N)
check('unit multiple', solve([N+1],1), N+1)
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
between multiples	`6`	`4`	Failed
already multiple	`3`	`2`	Failed
zero length	`1`	`0`	Failed
empty	`0`	`0`	Passed
mixed maximum	`4`	`3`	Failed
unit multiple	`3`	`2`	Failed

SHA-256 / d0b18764d490dc1164472b218bd05e6f7f1d5118137a7994b6e5156ee737ed7c

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 short integer-token lists under the stated toy contract; results do not claim compatibility with a specific training framework. 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 2026-09-30T16:48:08.895983+00:00.

Case digest / f4e40efc38fa442fa089935f16540e6df098df7460193d4b3a9bcdbcf06fbb50