what made local AI click for you on B200

[plaingarden18]

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 embeddings.

I'm especially interested in mistakes worth avoiding.

[cacheplane98]

A couple things I'd check first. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at token cost.

[bytestack]

worth checking the driver first.

[sparrowhall]

My first guess would be this: On 5090, the easiest win is often to back off the aggressive settings and see whether the problem survives.

[quiettoken]

My first guess would be this: On 5090, the easiest win is often to back off the aggressive settings and see whether the problem survives.

What exact setup are you running? Do you see the same thing with a smaller batch or shorter prompt?

[mangofield33]

Interesting, I hadn't tried that. I don't think your conclusion is far off. The biggest change for me was simplifying the stack instead of adding one more optimization layer.

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

[thetagrid15]

My first guess would be this: Before rebuilding anything, I'd test one clean baseline with Gemma and compare from there. Not glamorous advice, but repeatable baselines save a lot of time.