+- Inference (https://www.datasculpting.ai/inference)
+-- Forum: Generative AI (https://www.datasculpting.ai/inference/forumdisplay.php?fid=17)
+--- Forum: Video (https://www.datasculpting.ai/inference/forumdisplay.php?fid=70)
+--- Thread: vision-language models on consumer gear (/showthread.php?tid=870)
vision-language models on consumer gear - visionforge - 10-07-2026
Wanted to compare notes with people doing similar things.
I'm trying to decide whether to keep the current stack or change course. Main question is how people here are handling embeddings.
Not looking for the perfect answer, mostly what worked in real use.
Re: vision-language models on consumer gear - rasterhall - 10-07-2026
That matches what I was seeing. I ended up in about the same place using vLLM. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.
Re: vision-language models on consumer gear - plaingarden18 - 10-10-2026
Good point. The pattern around router policy has been pretty similar for me.
Re: vision-language models on consumer gear - paperrunner - 10-11-2026
that fixed it for me too.
Re: vision-language models on consumer gear - audioforge - 10-11-2026
Need a little more detail here. Do you see the same thing with a smaller batch or shorter prompt?
Re: vision-language models on consumer gear - quietstack - 10-11-2026
Interesting, I hadn't tried that. I don't think your conclusion is far off. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.
Re: vision-language models on consumer gear - latentpeak22 - 10-12-2026
Ran a quick check on my box. I couldn't reproduce the slowdown until I pushed context much higher.
Re: vision-language models on consumer gear - visionforge - 10-12-2026
A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at speculative decoding. Once you know the plain path works, then add the faster path back in piece by piece.