Module-Aware Optimization for Graph Neural Networks

媒体报道 Hugging Face Papers 原文语言:中文

论文摘要(作者报告,未经本站复测):Different optimizers impose different geometries on parameter updates.

全文

GraphMuon adapts Muon's matrix orthogonalization to the covariance of graph-propagated representations, reducing graph-induced update amplification and speeding validation-loss time-to-target on graph benchmarks.

正文以来源语言显示。

参考来源

← 返回每日AI报告