Very Short Responses Running External Model Weights With a Llamafile

[michaelgetachew-abebe]

Hello everyone,

I am a bit new to Llamafile. I was trying to run Llama2 using the 5 bit quantized gguf weights. However, I am experiencing very short responses is there a way I can adjust the response length in Llamafile? Or how can I make it respond with better and longer reponses?

Thanks in advance

[jart]

Could you copy and paste the command you're running and the output?

[pawel665j]

yes hi, I'm just stuck, I devote all my time to studying, very closely. thanks for the help. @echo off call .\llamafile.exe -m model-f16.gguf -t 52 -c 2048 -b 1024 [--languare-ru_RU] -ngl 9999 [--gpu-AUTO] -o log.txt [--save\all\logits-main.log] [--log-test] [--log-enable] [--log-append] [--interactive-first] pause

…

________________________________ From: Justine Tunney ***@***.***> Sent: Friday, November 29, 2024 10:57 PM To: Mozilla-Ocho/llamafile ***@***.***> Cc: Subscribed ***@***.***> Subject: Re: [Mozilla-Ocho/llamafile] Very Short Responses Running External Model Weights With a Llamafile (Discussion #645) Could you copy and paste the command you're running and the output? ― Reply to this email directly, view it on GitHub<#645 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AOOFRCBFHCAABKTX33QYXAL2DDBMDAVCNFSM6AAAAABSW3LSKKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTCNBRHA3TANA>. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>