Need a way to display prefill and decode metrics for geniex serve
Author: kmounika612-designCreated Sep 17, 2026Updated Sep 18, 2026
Is your feature request related to a problem? Please describe. Need a way to display prefill and decode metrics for geniex serve.
Describe the solution you'd like Just like we see the metrics for geniex infer and llama-server.exe, add prefill and decode t/s info for each request.
Describe alternatives you've considered N/A
Source: qualcomm/GenieX