speculative decoding in llama.cpp with SGLang

[mellowwire98]

Interesting, I hadn't tried that. The pattern around continuous batching has been pretty similar for me.

[kernelghost]

I'd approach it in a boring order. On A6000, the easiest win is often to back off the aggressive settings and see whether the problem survives. Once you know the plain path works, then add the faster path back in piece by piece.

[macroforge]

Appreciate the detail. 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.