Amazon AGI: super-selective checkpointing and resilient training integration

What changed

Scientific scope

Documented super-selective activation checkpointing, asynchronous checkpoint discovery and communication/allocator adjustments in a multimodal model-training system spanning Trainium and GPU clusters. Checkpointing, sharding and asynchronous restart mechanisms have predecessors. Inspected report gives system-level methods and self-reported efficiency, not a complete SSC kernel specification or independent replay. Family first release/citation year is2024, but this March2025 snapshot controls the acquired technical detail; no earlier SSC priority asserted.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Checkpointing, sharding and asynchronous restart mechanisms have predecessors. Inspected report gives system-level methods and self-reported efficiency, not a complete SSC kernel specification or independent replay. Family first release/citation year is2024, but this March2025 snapshot controls the acquired technical detail; no earlier SSC priority asserted.

Contributors

Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction