Open Models Almanac
Methodsevaluation

single mention · evaluation

Diverse benchmark validation for quantization

Validates quantization on a mixed benchmark set to detect unintended effects across task types.

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

Filed under evaluation