Distributed Training · 2018

Mixed Precision Training

Paulius Micikevicius, Sharan Narang, Jonah Alben, et al.

Showed that most neural-network operations can run in 16-bit floating point rather than 32-bit, roughly halving memory use and speeding training on tensor hardware without accuracy loss.

Editorial record

Plain-language summary

The authors stored and computed most values in FP16 while keeping a master copy of the weights in FP32 for stable updates, and introduced loss scaling to shift small gradient values into FP16's representable range. Combined with FP32 accumulation for reductions, this matched full-precision accuracy across image, speech, language, and GAN models. The technique cut memory footprint and doubled arithmetic throughput on hardware with 16-bit units, becoming a default ingredient in large-scale training.

Knowledge graph

Relationships

Antecedents

Source record

Provenance

Record ID
P-130
Record created
2026-07-13
Last reviewed
2026-07-14
Record version
2

Citation caveat: Citation metadata is approximate and marked unverified in the source dataset.