@robi said in Serge - LLaMa made easy 🦙 - self-hosted AI Chat:
@lars1134 the tests I've run in a LAMP app allows these to run in less than 5GB RAM (depending on model) using the right combo of sw all on CPU only.
Awesome, I will definitely look into that today and tomorrow. Thank you for the inspiration 
. I will let you know how things went for me.
@Kubernetes Thank you for looking into it. I understand the limitations. I run a few servers that have more than 10 EPYC cores and 50GB RAM left unused - those servers have a lot more available resources than our local clients. But I understand that not too many have a similar situation.
It looks like that only runs locally, but I like the program.
I liked serge as I have quite a bit of ram left unused at one of my servers.
Is there any interest in getting this in the app store?
@nebulon thank you for the incredibly fast update and the info 
I was quite confused at my first attempt at posting this. There was an error that I was not privileged. I didn't know what to do 
Anyhow, thanks!
Hello there, 1.18 released a couple of days ago. Is there a possibility to update the package?
I cant seem to figure out how this forum works or how to get privilages. I am a paying subscriber, Do I need to link accounts or something?
