Embeddings discussion

[vectorshed43]

Did a bit of testing and thought I'd post the result here.

I tried this with Phi using PyTorch on 5090. The headline result was better than I expected, but batch size still ended up being the limiting factor.

Curious if anyone else landed in the same place.

My use case is pretty ordinary: local chat, some document work, and occasional batch jobs.

[paperport64]

That matches what I was seeing. I don't think your conclusion is far off. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

[edgecoast]

That matches what I was seeing. I don't think your conclusion is far off. The biggest change for me was simplifying the stack instead of adding one mo

Good point. That matches what I saw on B200. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[vectorshed43]

A couple things I'd check first. On mini PC, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[mangoquill]

A couple questions before I guess too hard. Are you testing this on 5090 or something else?

[ggufwire22]

Can you fill in a bit more context? Are you testing this on dual 3090 or something else?

[vectorshed43]

Can you fill in a bit more context? Are you testing this on dual 3090 or something else?

What worked for me was pretty simple. If you're on LM Studio, I'd confirm the version first and then reduce variables until the behavior is obvious. Once you know the plain path works, then add the faster path back in piece by piece.

[quantnest]

Can you fill in a bit more context? Are you testing this on dual 3090 or something else?

Did a small comparison this afternoon. I couldn't reproduce the slowdown until I pushed context much higher. I expected the model choice to matter more than it did.

[vectorshed43]

Can you fill in a bit more context? Are you testing this on dual 3090 or something else?

What worked for me was pretty simple. On B200, the easiest win is often to back off the aggressive settings and see whether the problem survives. Not glamorous advice, but repeatable baselines save a lot of time.

[rasterhall]

What worked for me was pretty simple. If you're on vLLM, 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.