single mention · inference & serving
Linear E2M1 residual correction
A residual correction pass using a linear E2M1 residual for FP4 weight quantization.
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 linear E2M1 residual.
Filed under inference & serving › inference quantization