FA-97971 / Training label alignment / Open access

Class weights are emitted in class order instead of example order · case 01

Class weights are emitted in class order instead of example order.

Verified by executionVariant 1 · 7 checks per implementationDownload source bundle ↓JSON ↗

ROOT CAUSE

The routine enumerates class weights rather than looking up one weight for each example in sequence.

THE FAILURE

The routine enumerates class weights rather than looking up one weight for each example in sequence.

Unsuccessful approach: Sorting examples before lookup returns valid weights in the wrong row order.

Case contract

Return one weight per label from the supplied class-to-weight mapping; unknown labels use the supplied default.

Why this case matters

A small deterministic training-data operation makes label ownership and ordering explicit.

1 / The failure

Exit 1

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

N = 1
observations = []
def solve(labels, weights, default):
    return [weights[k] for k in sorted(weights) if k in labels]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('input order retained',solve(['z','a'],{'z':N,'a':N+1},7),[N,N+1])
check('unknown default',solve(['q'],{'a':1},N),[N])
check('empty labels',solve([],{ 'a':N},0),[])
check('zero weight',solve(['x'],{'x':0},9),[0])
check('repeated class',solve(['b','a','b'],{'a':N,'b':N+1},0),[N+1,N,N+1])
check('all unknown',solve(['u','v'],{},N),[N,N])
check('single known',solve(['a'],{'a':N},0),[N])
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
input order retained	`[2, 1]`	`[1, 2]`	Failed
unknown default	`[]`	`[1]`	Failed
empty labels	`[]`	`[]`	Passed
zero weight	`[0]`	`[0]`	Passed
repeated class	`[1, 2]`	`[2, 1, 2]`	Failed
all unknown	`[]`	`[1, 1]`	Failed
single known	`[1]`	`[1]`	Passed

SHA-256 / 2b346492692ab827f8e18b586aa3dd9d7d06a3b9c9262da4a398e3a0f87ffaac

2 / The unsuccessful fix

Exit 1

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

N = 1
observations = []
def solve(labels, weights, default):
    return [weights.get(k,default) for k in sorted(labels)]
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
check('input order retained',solve(['z','a'],{'z':N,'a':N+1},7),[N,N+1])
check('unknown default',solve(['q'],{'a':1},N),[N])
check('empty labels',solve([],{ 'a':N},0),[])
check('zero weight',solve(['x'],{'x':0},9),[0])
check('repeated class',solve(['b','a','b'],{'a':N,'b':N+1},0),[N+1,N,N+1])
check('all unknown',solve(['u','v'],{},N),[N,N])
check('single known',solve(['a'],{'a':N},0),[N])
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
input order retained	`[2, 1]`	`[1, 2]`	Failed
unknown default	`[1]`	`[1]`	Passed
empty labels	`[]`	`[]`	Passed
zero weight	`[0]`	`[0]`	Passed
repeated class	`[1, 2, 2]`	`[2, 1, 2]`	Failed
all unknown	`[1, 1]`	`[1, 1]`	Passed
single known	`[1]`	`[1]`	Passed

SHA-256 / 1899d3466d306d6a221a47a56dbdb75fdd16782a43abe6bb246b5431ed432fc3

HELD IN THE MEMBER ARCHIVE

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

This mechanism has 7 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

Class keys are unique dictionary keys; weights are finite scalars. 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 2026-09-30T16:48:07.975749+00:00.

Case digest / 027c015e20f238983f24f2b413c4634b1abd01dbf32a1a0313bf05aae5fea337