single mention · inference & serving
AWQ INT4 weight-only quantization
AWQ-based INT4 weight quantization with 16-bit activations, calibrated on agentic trajectories.
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 INT4 AWQ weight quantization.
Filed under inference & serving › inference quantization