LocalLLaMA
Need help building a rig / estimating performance for big LLMs to run when fully offline
Hello. So, first, ill probably explain the context, and why we (well, I kinda have a team, so ill say we) want to build something like that. We live in Russia. There are many preconditions, that point to the fact that pretty soon, we'll get our internet access (global internet, apparently) completely cut off. And we are preparing to go fully offline for indefinite period. I know, this sounds weird. No one believes that. But I prefer to be ready for the worst case scenario. ...and we kinda wan...
Read the full discussion on Reddit
This article was aggregated from Reddit. Click to join the conversation.
View on Reddit