Sheared LLaMA: Accelerating Language Model Pre-Training via Structured Pruningxiamengzhou.github.io3 points·jmorgan··0 commentsOpen articleSaveView on HN