技术分享·
Transformer(6) - LayerNorm
拿一个四维向量手算 LayerNorm:按哪一维统计,减均值、除标准差各在做什么,γ 和 β 怎么用,以及它在 Transformer 里的位置。

— code, plants, and everything in between
Senior frontend engineer specializing in large-scale application architecture and full-lifecycle performance optimization. From Xi'an, Shaanxi. Career spans collaborative tools, Web3, and low-code platforms. Avid gardener.
Thoughts & tutorials from the field
拿一个四维向量手算 LayerNorm:按哪一维统计,减均值、除标准差各在做什么,γ 和 β 怎么用,以及它在 Transformer 里的位置。