This program works pretty well! I can see the Spark realtime decode speed so far, but is there any way to see the cached prefill and uncached prefill speeds? I'd like to record these metrics on a daily basis to track the token generation status of my machine.
This program works pretty well! I can see the Spark realtime decode speed so far, but is there any way to see the cached prefill and uncached prefill speeds? I'd like to record these metrics on a daily basis to track the token generation status of my machine.