[从0开始AIGC][LLM]:Pre-Norm or Post-Norm?训练效率还是训练效果?
Pre-NormorPost-NormPre-NormorPost-Norm1.什么是Pre-Norm和Post-Norm2.为什么Pre-Norm比Post-Norm易于训练2.1Transformer:Attentionisallyourneed-PostNorm2.2Pre-Norm的提出:TransformerswithoutTears:ImprovingtheNormalizationo