Hugging Face's logo

TheBloke
/
LLongMA-2-7B-GGML

Transformers
llama
text-generation-inference
Model card Files Community
2

This model LLongMA-2-7B ha 8K contest?

#2
by mirek190 - opened
Discussion
mirek190

Do I properly understand it?
If yes then under llama.cpp I just need --ctx-size 8186 ?

Nexesenex

I'd say use
-c 8192 --rope-freq-base 10000 --rope-freq-scale 0.5"
As TheBloke wrote.

mirek190

Ohh I see right now ... thanks

Sign up or log in to comment