FA-99269 / Distributed training aggregation / Member archive

Gradient shards are clipped independently rather than by the global norm · case 04

Gradient shards are clipped independently rather than by the global norm.

Member previewVariant 4 · 3 implementations · 7 checks per implementation

Case contract

Given disjoint gradient shards and a nonnegative limit, return the same shard layout scaled by min(1, limit/global_L2_norm); zero norm uses scale one and results are rounded to 12 decimals.

Why this case matters

A toy distributed-training contract makes replica and shard contribution rules explicit without claiming framework compatibility or performance.

One recorded failure

Sample boundary fixture

This sample comes from the broken implementation of a controlled reproducer.

Boundary fixture	Actual	Expected	Outcome
norm spans two shards	`[[8.0, 0.0], [0.0, 8.0]]`	`[[4.8, 0], [0, 6.4]]`	Failed

MEMBER ARCHIVE

The complete case is available to members.

This record includes three runnable implementations, regression fixtures, execution results, and source hashes.

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