XDA Developers on MSN

My 8GB GPU shouldn't run flagship local LLMs, but this workflow makes it work anyway

I'm not shopping for a new GPU just yet ...