ParentFull threadgaleos·These are MLPerf training results. I think current ternary quantization research is focused more on speeding up inference?View on HN