This is promising: you can run LLM inference and training on 13W of power. I’ve yet to read the research paper, but they found you don’t need matrix multiplication if you adopt ternary [-1, 0, 1] values.
This is promising: you can run LLM inference and training on 13W of power. I’ve yet to read the research paper, but they found you don’t need matrix multiplication if you adopt ternary [-1, 0, 1] values.