jeremy ardley wrote: > Ask ChatGPT4 . Explain what you are trying to do and get it > to give you a suitable context and prompt
I don't know what to ask exactly, maybe I can ask ChatGPT4 ... > localdocs contains text you trust that can be used in > responses in preference to something synthesised from > the internet. I tried to google but it seems everything is local with the llamafile so I don't know, also if it was included it would have been in the man page or appeared as a simple plugin when you googled it. Anyway, the context is big enough to play around with for now then. Or the context size is maybe the same? Just when looking for an answer, the AI first looks for the localdocs, then the internet? Or you can throw in stuff there, and this gets crunched by the AI? But then, I don't have that much stuff, so I would have to get it from the internet anyway. Wait, let's first find out how big a file you can use! In my command, we see that -c or --ctx-size "Set the size of the prompt context" (llamafile(1)). If the unit is tokens, that means I know have not 512 but 2048 of them. There are a lot of token related stuff in the log even when it is disabled, seemingly. --prompt-cache-all will maybe store some data on the file or when you exit the application. Then that sweat stuff will be brought just right in again. taskset -c 0-2 \ ./llm \ --cli \ --log-disable \ --prompt-cache-all \ --silent-prompt \ -c 2048 \ -ngl 9999 \ -p "$(cat $src)" > $dst -- underground experts united https://dataswamp.org/~incal
