Pre-training LLM without Learning Rate Decay Enhances Supervised Fine-Tuning

• URL: http://arxiv.org/abs/2603.16127v1
• Date: Tue, 17 Mar 2026 05:17:07 GMT
• Title: Pre-training LLM without Learning Rate Decay Enhances Supervised Fine-Tuning
• Abstract summary: We investigate the role of learning rate scheduling in the large-scale pre-training of large language models.<n>We show that Warmup-Stable-Only (WSO) maintains a constant learning rate after warmup without any decay.
• Score: 17.543161067319023
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We investigate the role of learning rate scheduling in the large-scale pre-training of large language models, focusing on its influence on downstream performance after supervised fine-tuning (SFT). Decay-based learning rate schedulers are widely used to minimize pre-training loss. However, despite their widespread use, how these schedulers affect performance after SFT remains underexplored. In this paper, we examine Warmup-Stable-Only (WSO), which maintains a constant learning rate after warmup without any decay. Through experiments with 1B and 8B parameter models, we show that WSO consistently outperforms decay-based schedulers in terms of performance after SFT, even though decay-based schedulers may exhibit better performance after pre-training. The result also holds across different regimes with mid-training and over-training. Loss landscape analysis further reveals that decay-based schedulers lead models into sharper minima, whereas WSO preserves flatter minima that support adaptability. These findings indicate that applying LR decay to improve pre-training metrics may compromise downstream adaptability. Our work also provides practical guidance for training and model release strategies, highlighting that pre-training models with WSO enhances their adaptability for downstream tasks.

Related papers

• (How) Learning Rates Regulate Catastrophic Overtraining [31.988459260329353]
  Supervised fine-tuning (SFT) is a common first stage of LLM post-training.<n>We investigate catastrophic forgetting in finetuning through the lens of implicit regularization of the learning rate.<n>We link forgetting to overtraining: learning rate decay increases the sharpness of the pretrained model, which in turn exacerbates catastrophic forgetting during SFT.
  arXiv  Detail & Related papers  (2026-04-15T08:53:42Z)
• Amortized Predictability-aware Training Framework for Time Series Forecasting and Classification [10.816479922364097]
  We propose a general Amortized Predictability-aware Training Framework (APTF) for both time series forecasting (TSF) and time series classification (TSC)<n>APTF introduces two key designs that enable the model to focus on high-predictability samples while still learning appropriately from low-predictability ones.
  arXiv  Detail & Related papers  (2026-02-18T06:59:05Z)
• Class Confidence Aware Reweighting for Long Tailed Learning [0.8297806372438926]
  We present the design of a class and confidence-aware re-weighting scheme for long-tailed learning.<n>We use an (p_t, f_c) function to enable the modulation of the contribution towards the training task based upon the confidence value of the prediction.
  arXiv  Detail & Related papers  (2026-01-22T12:58:05Z)
• CurES: From Gradient Analysis to Efficient Curriculum Learning for Reasoning LLMs [53.749193998004166]
  Curriculum learning plays a crucial role in enhancing the training efficiency of large language models.<n>We propose CurES, an efficient training method that accelerates convergence and employs Bayesian posterior estimation to minimize computational overhead.
  arXiv  Detail & Related papers  (2025-10-01T15:41:27Z)
• Functional Scaling Laws in Kernel Regression: Loss Dynamics and Learning Rate Schedules [9.332823269318842]
  Scaling laws have emerged as a unifying lens for understanding and guiding the training of large language models.<n>We establish a Functional Scaling Law that captures the full loss trajectory under arbitrary LRSs.<n>We derive explicit scaling relations in both data- and compute-limited regimes.
  arXiv  Detail & Related papers  (2025-09-23T16:05:16Z)
• WSM: Decay-Free Learning Rate Schedule via Checkpoint Merging for LLM Pre-training [64.0932926819307]
  We present Warmup-Stable and Merge (WSM), a framework that establishes a formal connection between learning rate decay and model merging.<n>WSM provides a unified theoretical foundation for emulating various decay strategies.<n>Our framework consistently outperforms the widely-adopted Warmup-Stable-Decay (WSD) approach across multiple benchmarks.
  arXiv  Detail & Related papers  (2025-07-23T16:02:06Z)
• Can Pre-training Indicators Reliably Predict Fine-tuning Outcomes of LLMs? [42.608899417822656]
  We construct a dataset using 50 1B parameter LLM variants with systematically varied pre-training configurations.<n>We introduce novel unsupervised and supervised proxy metrics derived from pre-training that successfully reduce the relative performance prediction error rate by over 50%.
  arXiv  Detail & Related papers  (2025-04-16T21:19:09Z)
• A Multi-Power Law for Loss Curve Prediction Across Learning Rate Schedules [67.87680482844884]
  We present an empirical law that describes how the pretraining loss of large language models evolves under different learning rate schedules.<n>Our proposed law takes a multi-power form, combining a power law based on the sum of learning rates and additional power laws to account for a loss reduction effect induced by learning rate decay.
  arXiv  Detail & Related papers  (2025-03-17T04:36:45Z)
• The Journey Matters: Average Parameter Count over Pre-training Unifies Sparse and Dense Scaling Laws [51.608402959163925]
  We present the first systematic exploration of optimal sparse pre-training configurations for large language models.<n>We find that initiating pruning at 25% of total training compute and concluding at 75% achieves near-optimal final evaluation loss.<n>We propose a new scaling law that modifies the Chinchilla scaling law to use the average parameter count over pre-training.
  arXiv  Detail & Related papers  (2025-01-21T20:23:22Z)
• Overcoming Recency Bias of Normalization Statistics in Continual Learning: Balance and Adaptation [67.77048565738728]
  Continual learning involves learning a sequence of tasks and balancing their knowledge appropriately. We propose Adaptive Balance of BN (AdaB$2$N), which incorporates appropriately a Bayesian-based strategy to adapt task-wise contributions. Our approach achieves significant performance gains across a wide range of benchmarks.
  arXiv  Detail & Related papers  (2023-10-13T04:50:40Z)
• Extrapolation for Large-batch Training in Deep Learning [72.61259487233214]
  We show that a host of variations can be covered in a unified framework that we propose. We prove the convergence of this novel scheme and rigorously evaluate its empirical performance on ResNet, LSTM, and Transformer.
  arXiv  Detail & Related papers  (2020-06-10T08:22:41Z)

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.