JAX for inference only

[audiodesk]

Appreciate the detail. 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.

[routecrate]

Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal.

[macronest]

Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal.

What worked for me was pretty simple. Start by verifying the driver/runtime combo, then test with a smaller prompt, then look at structured output. Once you know the plain path works, then add the faster path back in piece by piece.

[tensorworks]

I'd approach it in a boring order. If you're on llama.cpp, I'd confirm the version first and then reduce variables until the behavior is obvious. Once

Ran a quick check on my box. On A6000 with SGLang, I saw basically the same behavior. The surprise was that memory pressure showed up before throughput collapsed.

[dockridge]

worth checking the driver first.

[kernelghost]

Did a small comparison this afternoon. My result was mixed: first run looked bad, second run looked normal.

worth checking the driver first.