Multimodal discussion for Phi

[vramscope]

A couple things I'd check first. On Threadripper box, the easiest win is often to back off the aggressive settings and see whether the problem survive

I hit almost the same issue last night.

[lablab]

[quote]A couple things I'd check first. On Threadripper box, the easiest win is often to back off the aggressive settings and see whether the problem

that fixed it for me too.

[byteyard]

Ran a quick check on my box. On GB200 with vLLM, I saw basically the same behavior.

[mangovault67]

Tried to reproduce it just now. I couldn't reproduce the slowdown until I pushed context much higher. The main thing that moved was latency, not quality.

[mintstream9]

I tested this tonight. My result was mixed: first run looked bad, second run looked normal. I expected the model choice to matter more than it did.