3 ms·Paper summary: LLM training and inference without matrix multiplications2 points by calebgcc 2y ago