whisper alternatives lately with TensorRT-LLM

[fusedhall]

This keeps coming up in conversation so I figured I'd start a thread.

One thing I keep noticing is that people underestimate MoE serving until they actually try to run this stuff for more than one afternoon.

Curious whether that lines up with what other people are seeing.

I care more about stability than squeezing the last possible tokens/sec out of it.

[amberpatch25]

Good point. That matches what I saw on M3 Max. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[tensorworks]

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.

[vramarch]

That matches what I was seeing. I don't think your conclusion is far off. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[cudascope]

What exact setup are you running? Are you testing this on Threadripper box or something else?

[bytegarden73]

I tested this tonight. The issue went away once I changed one setting and restarted the stack cleanly. The main thing that moved was latency, not quality.

[visionforge]

What worked for me was pretty simple. On B200, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lot of these threads end up being one bad config plus a pile of conclusions drawn from it.

[oxidegrove40]

What worked for me was pretty simple. On B200, the easiest win is often to back off the aggressive settings and see whether the problem survives. A lo

I don't think that's the whole story. I get why that seems right, but on mini PC the bottleneck felt elsewhere. Could be a workload difference, but I wouldn't generalize too fast from one setup.

[bytegarden73]

Interesting, I hadn't tried that. The pattern around KV cache pressure has been pretty similar for me. Worth rechecking your defaults though, because one bad assumption can make the rest of the thread look confusing.

[cobaltworks25]

worth checking the driver first.