There are several ways to approach GPUs for Local AI, so I wanted to compare experiences.
I am particularly interested in learning curve and reliability. I would like to compare approaches that people have actually used rather than just theoretical suggestions.
What trade-offs have you found...