Search / Retrieval / Reranking discussion for Llama

[northgrid33]

Posting this in case it saves someone else a little time.

I'm helping a small team set up a local workflow. We don't need a giant stack, but we do care about speculative decoding.

Simple and boring is a valid answer.

[prompttinker44]

Interesting, I hadn't tried that. I ended up in about the same place using SGLang. Worth rechecking your defaults though, because one bad assumption c

What worked for me was pretty simple. 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.

[macrolab]

[quote]Interesting, I hadn't tried that. I ended up in about the same place using SGLang. Worth rechecking your defaults though, because one bad assum

My first guess would be this: Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at reranking. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[cudaworks]

Posting this in case it saves someone else a little time.

That matches what I was seeing. I ended up in about the same place using SGLang. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

[northcraft90]

Good point. The pattern around batch size has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[llm_wren]

Interesting, I hadn't tried that. I ended up in about the same place using SGLang. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.