Open Models Almanac
Methodsinference & servinginference quantization

single mention · inference & serving

Offline weight-layout permutation

A one-time offline permutation of weight layout used to reduce runtime dequantization overhead.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See inference quantization for the methods filed beside it.

Also called weight layout is permuted offline.

Filed under inference & serving › inference quantization