10-16-2026, 10:47 AM
I tested this tonight. On A6000 with PyTorch, I saw basically the same behavior. The main thing that moved was latency, not quality.
|
NVIDIA driver regression with vLLM
|
|
« Next Oldest | Next Newest »
|
| Messages In This Thread |
|
NVIDIA driver regression with vLLM - by audiopath66 - 10-05-2026, 04:17 AM
Re: NVIDIA driver regression with vLLM - by mintforge - 10-05-2026, 12:24 PM
Re: NVIDIA driver regression with vLLM - by orbitbench6 - 10-05-2026, 01:03 PM
Re: NVIDIA driver regression with vLLM - by phaseport36 - 10-16-2026, 11:20 AM
Re: NVIDIA driver regression with vLLM - by paperport64 - 10-07-2026, 01:42 AM
Re: NVIDIA driver regression with vLLM - by phaseport36 - 10-10-2026, 01:56 AM
Re: NVIDIA driver regression with vLLM - by dockember - 10-10-2026, 10:22 AM
Re: NVIDIA driver regression with vLLM - by relaynest82 - 10-13-2026, 10:47 AM
Re: NVIDIA driver regression with vLLM - by mintforge - 10-16-2026, 02:11 PM
Re: NVIDIA driver regression with vLLM - by cobaltspark - 10-16-2026, 10:47 AM
Re: NVIDIA driver regression with vLLM - by audiopath66 - 10-16-2026, 12:01 PM
Re: NVIDIA driver regression with vLLM - by sparrowkite53 - 10-18-2026, 02:54 AM
|
Users browsing this thread: 1 Guest(s)