Open Models Almanac
Methodsinference & servinginference quantization

single mention · inference & serving

Rotation or higher precision for Mamba input activations

Rotation or increased activation precision is used to mitigate effects of FP4 activation 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.

Filed under inference & serving › inference quantization