🚩 Report
#11
by
kshabana
- opened
I downloaded the lama 3 8b 8bit in LM studio form you and it was broken when i ask the model anything like hi it give me endless answers
kshabana
changed discussion status to
closed
kshabana
changed discussion status to
open
kshabana
changed discussion status to
closed
kshabana
changed discussion status to
open
Is this the base or instruct?
Instruct
Just tried the quants with lmstudio 0.2.20 (latest as of 23 Apr 24) on Mac, Im not able to reproduce the issue. Can you please recheck if you downloaded QuantFactory/Meta-Llama-3-8B-Instruct-GGUF
What is the quant you're using?
i got this weird thing on m2 max
tried l3_instruct_q8 and l3_instruct_q8
but i have the lm studio 0.2.20
and i have the model that you use
Preset should be LLaMA 3 on the right
the preset is llama 3