Tag
The author shares their experience using Tesla P100 GPUs for local AI inference, finding them cost-effective and performant with optimizations via llama.cpp, despite initial advice from frontier models.