Anyone else unable to run gpt oss on their GPU? Why is it always defaulting to the CPU??
>downloaded the wrong llamacpp
No I didn't, if I load another model it loads on the GPU just fine.
I'm using these parameters
llama-server.exe ^
-m %MODEL% ^
-t 12 ^
-c 16384 ^
-fa ^
-np 8 ^
-ngl 65 ^
-v ^
--port 5001 ^
--host 0.0.0.0