Inference › Training & Customization › Distillation v
when RAG was overkill

Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Linear Mode
when RAG was overkill
tensorkite Offline
Member
***
Posts: 53
Threads: 4
Joined: Sep 2026
Reputation: 0
#8
10-18-2026, 05:05 PM
Quote:Interesting, I hadn't tried that. I ended up in about the same place using llama.cpp. I got better results once I stopped chasing one metric and just

I tested this tonight. On 7900 XTX with SGLang, I saw basically the same behavior.
Find
Reply


Messages In This Thread
when RAG was overkill - by homelabhenry - 10-12-2026, 11:39 PM
Re: when RAG was overkill - by quantshed - 10-14-2026, 11:43 AM
Re: when RAG was overkill - by sanescope - 10-15-2026, 11:49 AM
Re: when RAG was overkill - by southspark - 10-17-2026, 12:20 AM
Re: when RAG was overkill - by deepseeked - 10-17-2026, 03:10 AM
Re: when RAG was overkill - by plainscope50 - 10-17-2026, 04:28 AM
Re: when RAG was overkill - by fastbench - 10-17-2026, 03:27 AM
Re: when RAG was overkill - by tensorkite - 10-18-2026, 05:05 PM
Re: when RAG was overkill - by quiettoken - 10-18-2026, 08:41 PM
Re: when RAG was overkill - by papermixer4 - 10-18-2026, 11:22 PM
Re: when RAG was overkill - by turboarch95 - 10-18-2026, 11:50 PM
Re: when RAG was overkill - by mangohouse - 10-19-2026, 12:02 AM
Re: when RAG was overkill - by cobalttinker - 10-19-2026, 12:30 AM
Re: when RAG was overkill - by emberrunner - 10-19-2026, 01:05 AM
Re: when RAG was overkill - by thetaember38 - 10-19-2026, 03:12 AM
Re: when RAG was overkill - by rocmriley - 10-19-2026, 04:27 AM

Forum Jump:


Users browsing this thread: 1 Guest(s)