How to Set the Learning Rate for Large-Scale Pre-training?

• URL: http://arxiv.org/abs/2601.05049v1
• Date: Thu, 08 Jan 2026 15:55:13 GMT
• Title: How to Set the Learning Rate for Large-Scale Pre-training?
• Authors: Yunhua Zhou, Shuhao Xing, Junhao Huang, Xipeng Qiu, Qipeng Guo,
• Abstract summary: We formalize this investigation into two distinct research paradigms: Fitting and Transfer.<n>Within the Fitting Paradigm, we introduce a Scaling Law for search factor, effectively reducing the search complexity from O(n3) to O(n*C_D*C_) via predictive modeling.<n>We extend the principles of $$Transfer to the Mixture of Experts (MoE) architecture, broadening its applicability to encompass model depth, weight decay, and token horizons.
• Score: 73.03133634525635
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Optimal configuration of the learning rate (LR) is a fundamental yet formidable challenge in large-scale pre-training. Given the stringent trade-off between training costs and model performance, the pivotal question is whether the optimal LR can be accurately extrapolated from low-cost experiments. In this paper, we formalize this investigation into two distinct research paradigms: Fitting and Transfer. Within the Fitting Paradigm, we innovatively introduce a Scaling Law for search factor, effectively reducing the search complexity from O(n^3) to O(n*C_D*C_η) via predictive modeling. Within the Transfer Paradigm, we extend the principles of $μ$Transfer to the Mixture of Experts (MoE) architecture, broadening its applicability to encompass model depth, weight decay, and token horizons. By pushing the boundaries of existing hyperparameter research in terms of scale, we conduct a comprehensive comparison between these two paradigms. Our empirical results challenge the scalability of the widely adopted $μ$ Transfer in large-scale pre-training scenarios. Furthermore, we provide a rigorous analysis through the dual lenses of training stability and feature learning to elucidate the underlying reasons why module-wise parameter tuning underperforms in large-scale settings. This work offers systematic practical guidelines and a fresh theoretical perspective for optimizing industrial-level pre-training.

Related papers

• Answer First, Reason Later: Aligning Search Relevance via Mode-Balanced Reinforcement Learning [7.006180736433431]
  Building a search relevance model that achieves both low latency and high performance is a long-standing challenge in the search industry.<n>We propose a novel textbfAnswer-First, Reason Later (AFRL) paradigm.<n>This paradigm requires the model to output the definitive relevance score in the very first token, followed by a structured logical explanation.
  arXiv  Detail & Related papers  (2026-02-10T17:28:12Z)
• A Theoretical Framework for Auxiliary-Loss-Free Load Balancing of Sparse Mixture-of-Experts in Large-Scale AI Models [3.0247776995428945]
  In large-scale AI training, Sparse Mixture-of-Experts (s-MoE) layers enable scaling by activating only a small subset of experts per token.<n>We provide a theoretical framework for analyzing the Auxiliary-Loss-Free Load Balancing (ALF-LB) procedure.
  arXiv  Detail & Related papers  (2025-12-03T16:00:02Z)
• Scaling Laws and In-Context Learning: A Unified Theoretical Framework [0.0]
  In-context learning (ICL) enables large language models to adapt to new tasks from demonstrations without parameter updates.<n>We present a unified theoretical framework connecting scaling laws to ICL emergence in transformers.<n>We show that ICL performance follows power-law relationships with model depth $L$, width $d$, context length $k$, and training data $D$, with exponents determined by task structure.
  arXiv  Detail & Related papers  (2025-11-09T05:19:14Z)
• LARES: Latent Reasoning for Sequential Recommendation [96.26996622771593]
  We present LARES, a novel and scalable LAtent REasoning framework for Sequential recommendation.<n>Our proposed approach employs a recurrent architecture that allows flexible expansion of reasoning depth without increasing parameter complexity.<n>Our framework exhibits seamless compatibility with existing advanced models, further improving their recommendation performance.
  arXiv  Detail & Related papers  (2025-05-22T16:22:54Z)
• Supervised Optimism Correction: Be Confident When LLMs Are Sure [91.7459076316849]
  We establish a novel theoretical connection between supervised fine-tuning and offline reinforcement learning.<n>We show that the widely used beam search method suffers from unacceptable over-optimism.<n>We propose Supervised Optimism Correction, which introduces a simple yet effective auxiliary loss for token-level $Q$-value estimations.
  arXiv  Detail & Related papers  (2025-04-10T07:50:03Z)
• LESA: Learnable LLM Layer Scaling-Up [57.0510934286449]
  Training Large Language Models (LLMs) from scratch requires immense computational resources, making it prohibitively expensive.<n>Model scaling-up offers a promising solution by leveraging the parameters of smaller models to create larger ones.<n>We propose textbfLESA, a novel learnable method for depth scaling-up.
  arXiv  Detail & Related papers  (2025-02-19T14:58:48Z)
• Provable Meta-Learning with Low-Rank Adaptations [37.120226706944926]
  We introduce a framework for generic PEFT-based meta-learning to learn a model that can easily adapt to unseen tasks.<n>For linear models using LoRA, we show that standard retraining is provably suboptimal for finding an adaptable set of parameters.<n>We verify these theoretical insights through experiments on synthetic data as well as real-data vision and language tasks.
  arXiv  Detail & Related papers  (2024-10-29T17:24:18Z)
• Optimization Hyper-parameter Laws for Large Language Models [52.49860340549727]
  We present Opt-Laws, a framework that captures the relationship between hyper- parameters and training outcomes.<n>Our validation across diverse model sizes and data scales demonstrates Opt-Laws' ability to accurately predict training loss.<n>This approach significantly reduces computational costs while enhancing overall model performance.
  arXiv  Detail & Related papers  (2024-09-07T09:37:19Z)
• When Parameter-efficient Tuning Meets General-purpose Vision-language Models [65.19127815275307]
  PETAL revolutionizes the training process by requiring only 0.5% of the total parameters, achieved through a unique mode approximation technique. Our experiments reveal that PETAL not only outperforms current state-of-the-art methods in most scenarios but also surpasses full fine-tuning models in effectiveness.
  arXiv  Detail & Related papers  (2023-12-16T17:13:08Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.