Vector Databases discussion

[macroforge]

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

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

Nothing scientific here, just practical notes from a real setup.

The part that surprised me was how much the boring details mattered more than the model choice.

[routeforge64]

I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly.

[tinycontext]

What worked for me was pretty simple. If you're on SGLang, I'd confirm the version first and then reduce variables until the behavior is obvious. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

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

[orbitbench6]

I'd approach it in a boring order. If you're on PyTorch, 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.

[emberquill]

same here on my side.

[macronest]

Can you fill in a bit more context? Do you see the same thing with a smaller batch or shorter prompt?

[tinycontext]

My first guess would be this: If you're on TensorRT-LLM, 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.

[relayforge]

Ran a quick check on my box. My result was mixed: first run looked bad, second run looked normal.

[vectorshed43]

subscribed, curious where this lands.

[byteyard]

Good point. I ended up in about the same place using llama.cpp.