Open Models Almanac
Methodsinference & servinginference quantization

single mention · inference & serving

Quantization of the parallel-drafter language-model head

Additional quantization of the parallel drafter's LM head to improve throughput.

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 Further quantizing the LM head.

Filed under inference & serving › inference quantization