AWhat changed
Combined FP16 storage/computation, FP32 master updates and loss scaling to train tested neural networks with reduced activation bandwidth and memory while recovering higher-precision baselines.
Scientific scope
Combined FP16 storage/computation, FP32 master updates and loss scaling to train tested neural networks with reduced activation bandwidth and memory while recovering higher-precision baselines. Scaling must avoid overflow and gradients are unscaled before updates or clipping. FP32 updates/accumulation remain; reduced storage does not mean every operation or model trains entirely in FP16. Reported accuracy/resource outcomes are task- and implementation-specific.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Scaling must avoid overflow and gradients are unscaled before updates or clipping. FP32 updates/accumulation remain; reduced storage does not mean every operation or model trains entirely in FP16. Reported accuracy/resource outcomes are task- and implementation-specific.
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