Concept自适应归一化层
AdaRMSNorm
An RMSNorm normalization layer with adaptive modulation that dynamically adjusts normalization parameters according to conditional input. It is commonly used for multimodal or multi-morphology conditioning.
Timeline (last 90 days)
Oct 19
REKI pretraining preserves each dataset's native control convention, and each robot embodiment enters the action expert through a dedicated state MLP plus AdaRMSNorm.
Unverified50%