Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
prompt length getting silly
#7
Interesting, I hadn't tried that. The pattern around prefill/decode split has been pretty similar for me. The biggest change for me was simplifying the stack instead of adding one more optimization layer.
learning in public
Reply


Messages In This Thread
prompt length getting silly - by mlxgarden - 09-26-2026, 03:53 PM
Re: prompt length getting silly - by tensorworks - 09-27-2026, 12:22 AM
Re: prompt length getting silly - by batchember - 09-28-2026, 12:59 AM
Re: prompt length getting silly - by paperport64 - 09-28-2026, 01:26 AM
Re: prompt length getting silly - by mlxgarden - 09-28-2026, 01:30 PM
Re: prompt length getting silly - by mlxgarden - 09-30-2026, 05:00 PM
Re: prompt length getting silly - by cacheline - 09-29-2026, 02:03 AM
Re: prompt length getting silly - by mintstation99 - 09-30-2026, 02:13 PM
Re: prompt length getting silly - by quiettoken - 10-01-2026, 01:48 AM
Re: prompt length getting silly - by batchember - 10-01-2026, 10:15 AM
Re: prompt length getting silly - by rasterhall - 10-01-2026, 07:41 PM
Re: prompt length getting silly - by emberrunner - 10-01-2026, 10:08 PM

Forum Jump:


Users browsing this thread: 1 Guest(s)