The paper investigates why adaptive optimizers like Adam outperform SGD when fine‑tuning Transformers. It introduces gradient heterogeneity—the variation in gradient norms across parameter blocks—and shows, both theoretically and experimentally, that this heterogeneity, together with Hessian heterogeneity, hampers SGD convergence while sign‑based methods such as SignSGD are less affected. The study links the source of gradient heterogeneity to layer‑normalization placement, finding that Post‑LN architectures exhibit the strongest effect, and uses SignSGD as a tractable proxy to analyze Adam‑like behavior and learning‑rate scaling.
By Akiyoshi Tomihari, Issei Sato
arXiv:2607. 10593v1 Announce Type: new Abstract: Normalization is a critical component for stabilizing Transformer training, yet the choice between static strategies such as Layer Normalization (LN) and adaptive alternatives remains largely task-dependent.
By Piyush Kaushik Bhattacharyya, Divyanshu Rai, Swastik Singh, Kumar Aakash, Ayush Ranjan, Krutika Verma
arXiv:2510. 09904v2 Announce Type: replace-cross Abstract: Despite their widespread use, training deep Transformers can be unstable.
By Kelvin Kan, Xingjian Li, Benjamin J. Zhang, Tuhin Sahai, Stanley Osher, Krishna Kumar, Markos A. Katsoulakis
arXiv:2606. 31397v1 Announce Type: cross Abstract: State-based fine-tuning has emerged as a compelling alternative to weight-based adaptation for transformers, updating lightweight controls into states rather than model weights, offering substantial memory savings while retaining parameter efficiency.
By Duc Anh Nguyen, Tien Ngoc Luu, Tung Pham, Toan Tran
arXiv:2606. 00605v1 Announce Type: new Abstract: Transformers have achieved remarkable success across a wide range of applications, and a growing body of work suggests that part of their strength comes from their ability to learn and execute algorithmic procedures.
By Lyumin Wu, Chenyang Zhang, Yuan Cao
arXiv:2607. 19391v1 Announce Type: cross Abstract: Low-Rank Adaptation is widely used for parameter-efficient fine-tuning, yet existing methods typically assign the same adapter rank to every transformer layer despite their heterogeneous adaptation requirements.
By Ashutosh Tripathi, Surya Deep Singh, Pranab Sahoo, Sriparna Saha