Tag
A user reflects on the complexity and fascination of running local LLMs, touching on hardware selection, quantization, and tensor parallelism.