First time testing local text model I don’t know much yet.I’ve seen people with 8GB cards complaining that text generation is very slow so I don’t have much hope about that but still… I think I need to do some configuration, when generating text my SSD is at 100% reading 1~2gb/s while my GPU does not reach 15% usage.
Using RTX 2060 6GB, 16GB RAM.
This is the model I am testing ( mythomax-l2-13b.Q8_0.gguf): https://huggingface.co/TheBloke/MythoMax-L2-13B-GGUF/tree/main
So, you dont have enough ram to fit that model. It’s actually overrunning your ram entirely and using the wrong kind of vram, virtual ram, aka paged memory.
Idk what you’re trying to do but the best answer is openhermes 2.5 mistral 7B Q3 and 4k context or similar or maybe Rocket 3B Q6 would be even faster.
Hermes is king. I understand why you want that model, but 13bQ8 is huge, 17GBish memory at 8k context.
it will speed up if you get it off the hard drive at least, try a Q3k_l if you’re determined to run mythomax.
Never use the Q_8 versions of GGUFs unless most/all of the model can comfortably fit into your VRAM. The Q_6 version is much smaller, and almost the same quality.
For your setup, I would use mythomax-l2-13b.Q4_K_M.gguf.
In my tests Q4 is giving me the same amount of tokens as Q5 so I decided to use Q5, first time tesint text gen locally with models, thank you very much for explaining I am getting used to it now and understanding what the settings do.
Yeah there’s so much to learn I’m still figuring a lot out too.
Good tip for settings: Play around mostly with temperature, top-p, and min-p.