retrieval stack for private docs

[emberquill]

This keeps coming up in conversation so I figured I'd start a thread.

The more I work with local setups, the more it seems like reranking decides everything downstream.

Maybe obvious to some of you, but it clicked for me this week.

[macrolab]

I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up before throughput collapsed.

I'm happy to test one more variant if you want a second data point.

[vramghost]

I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly. The surprise was that memory pressure showed up

Appreciate the detail. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

There are a few different conversations hiding inside this thread, so being explicit about the goal helps.

[bytestack]

Interesting, I hadn't tried that. That matches what I saw on M3 Max.

[visionforge]

My first guess would be this: Before rebuilding anything, I'd test one clean baseline with DeepSeek and compare from there. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[embershed99]

Appreciate the detail. I ended up in about the same place using LM Studio. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[edgequill]

A couple things I'd check first. If you're on PyTorch, I'd confirm the version first and then reduce variables until the behavior is obvious. Not glamorous advice, but repeatable baselines save a lot of time.