whisper alternatives lately: setup advice

[visionforge]

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.

Would rather learn from your scars than collect my own all at once.

[latentpeak22]

Thanks, that helps. That matches what I saw on dual 3090.

[signalpeak]

That matches what I was seeing. I ended up in about the same place using llama.cpp. I got better results once I stopped chasing one metric and just looked at whether the workflow felt usable.

If you post your exact settings, people here can probably narrow it down pretty fast.

[bytegarden73]

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

[vramarch]

What exact setup are you running? Are you testing this on dual 3090 or something else?

[batchwire]

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at batch size.

[amberpatch25]

I hit almost the same issue last night.

[quiethouse]

following this because I'm in the same boat.

I think the bottleneck is somewhere else. I wouldn't optimize around that first. On my side the bigger swing came from the runtime and not the model. Worth testing the simpler explanation before adopting the more interesting one.

[kernelpeak]

Appreciate the detail. That matches what I saw on 5090.

[vramarch]

I'd approach it in a boring order. If you're on Ollama, I'd confirm the version first and then reduce variables until the behavior is obvious. Once you know the plain path works, then add the faster path back in piece by piece.