bartowski

/

nvidia_Nemotron-H-8B-Reasoning-128K-GGUF

like 2

Hi there, thanks a lot for generating this GGUF!
I'm trying to run this with llama-server, but it just generates and endless loop of gibberish tags. The same thing happens with bartowski/nvidia_NVIDIA-Nemotron-Nano-9B-v2-GGUF and bartowski/nvidia_NVIDIA-Nemotron-Nano-12B-v2-GGUF:

im_end_start<_enduserim>_start|imassim_start<ass userass_end>
ass|<>
hi
_starthiuseristant_end_start>

<
istanthihiass_startass>istant|<im<
im
im
_endistant_start>
<<ass_end>>
userhiassuseristant<hi_start>userhiim
imhi
_start_endistant_end
_endistant|istant_end_start>im<<
assistant_end
_endistant<
_start_starthi<>|istant<userim_endhi
istant>

    ass_end_enduserasshiassassim_end
    _start>
    |
    istant_startassim|_end_startass|hi<>
    ass>
    |_endistant<