09-30-2026, 01:58 AM
I tested this tonight. I couldn't reproduce the slowdown until I pushed context much higher.
|
CUDA update broke my throughput with Ollama
|
|
« Next Oldest | Next Newest »
|