稳定超网络学习的非比例参数化（英文版）

英文标题：Non-Proportional Parametrizations for Stable Hypernetwork Learning中文摘要：本文探讨了超网络训练中出现的不稳定性问题，分析发现这与常见的超网络模型架构和初始化方法存在关联，提出了一种基于非比例加性参数化的改进超网络构建方法，在多项任务中实验证明该方法能够更加稳定地提高超网络的训练速度和收敛性。英文摘要：Hypernetworks are neural networks that generate the parameters of anotherneural network. In many scenarios, current hypernetwork training strategies areunstable, and convergence is often far slower than for non-hypernetwork models.We show that this problem is linked to an issue that arises when using commonch