Beyond LoRA vs. Full Fine-Tuning: Gradient-Guided Optimizer Routing for LLM Adaptation

• URL: http://arxiv.org/abs/2605.07111v1
• Date: Fri, 08 May 2026 01:38:58 GMT
• Title: Beyond LoRA vs. Full Fine-Tuning: Gradient-Guided Optimizer Routing for LLM Adaptation
• Abstract summary: Full Fine-Tuning (FFT) provides the representational plasticity required for high-entropy knowledge injection.<n>Low-Rank Adaptation (LoRA) can match or surpass FFT performance because many tasks only require updates in a low-rank space and benefit from LoRA's additional regularization.<n>We propose a Mixture of LoRA and Full (MoLF) Fine-Tuning, a unified framework that enables continuous navigation between both training regimes.
• Score: 15.4865294569737
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Recent literature on fine-tuning Large Language Models highlights a fundamental debate. While Full Fine-Tuning (FFT) provides the representational plasticity required for high-entropy knowledge injection, Low-Rank Adaptation (LoRA) can match or surpass FFT performance because many tasks only require updates in a low-rank space and benefit from LoRA's additional regularization. Through empirical evaluation across diverse tasks (SQL, Medical QA, and Counterfactual Knowledge) and varying language models (Gemma-3-1B, Qwen2.5-1.5B, and Qwen2.5-3B), we verify both trends and demonstrate that relying solely on either static architecture is structurally limited. To address this challenge, we propose a Mixture of LoRA and Full (MoLF) Fine-Tuning, a unified framework that enables continuous navigation between both training regimes. MoLF dynamically routes updates between FFT and LoRA at the optimizer level to ensure that exact gradient signals are available to both experts throughout training, yielding stable training dynamics. For memory-constrained environments, we also introduce MoLF-Efficient, which freezes base weights and only routes updates among a pair of LoRA experts of potentially varying rank. Our evaluations show that MoLF either improves on or stays within $1.5\%$ of the better of FFT and LoRA across all settings, while MoLF-Efficient outperforms prior adaptive LoRA approaches by up to $20\%$ on Fact and $9\%$ on Med and SQL.

Related papers

• Taming Momentum: Rethinking Optimizer States Through Low-Rank Approximation [85.89510825889168]
  We introduce LoRA-Pre, a novel low-rank system for efficient pre-training.<n>LoRA-Pre decomposing the momentum matrix into a compact low-rank subspace within the online linear learner.<n>We empirically validate LoRA-Pre's efficacy by pre-training models from the Llama architecture family.
  arXiv  Detail & Related papers  (2026-02-27T18:57:06Z)
• Make LoRA Great Again: Boosting LoRA with Adaptive Singular Values and Mixture-of-Experts Optimization Alignment [20.382810396966473]
  Low-Rank Adaptation (LoRA) enables parameter-efficient fine-tuning for Large Language Models (LLMs)<n>Current methods optimize LoRA by initializing with static singular value decomposition subsets, leading to suboptimal leveraging of pre-trained knowledge.<n>We propose underlineGreat LunderlineoRunderlineA Mixture-of-Experunderlinet (GOAT)<n>GOAT integrates relevant priors using an SVD-structured MoE, and aligns optimization with full fine-tuned MoE by deriving a theoretical scaling factor
  arXiv  Detail & Related papers  (2025-02-24T06:48:13Z)
• BeamLoRA: Beam-Constraint Low-Rank Adaptation [51.52097743781401]
  Low-Rank Adaptation (LoRA) has been widely adopted as one of the most effective parameter-efficient fine-tuning methods.<n>We propose BeamLoRA, which conceptualizes each LoRA module as a beam where each rank naturally corresponds to a potential sub-solution.
  arXiv  Detail & Related papers  (2025-02-19T10:33:22Z)
• Adaptive Parameter-Efficient Federated Fine-Tuning on Heterogeneous Devices [24.725928966071212]
  Federated fine-tuning (FedFT) has been proposed to fine-tune the pre-trained language models in a distributed manner.<n>We propose a novel LoRA-based FedFT framework, termed LEGEND, which faces the difficulty of determining the number of LoRA layers.<n>We analyze the coupled relationship between LoRA depth and rank distribution, and design an efficient LoRA configuration algorithm for heterogeneous devices.
  arXiv  Detail & Related papers  (2024-12-28T04:00:42Z)
• Unlocking Tuning-Free Few-Shot Adaptability in Visual Foundation Models by Recycling Pre-Tuned LoRAs [76.40876036912537]
  Large Language Models (LLMs) demonstrate strong few-shot adaptability without requiring fine-tuning.<n>Current Visual Foundation Models (VFMs) require explicit fine-tuning with sufficient tuning data.<n>We propose a framework, LoRA Recycle, that distills a meta-LoRA from diverse pre-tuned LoRAs with a meta-learning objective.
  arXiv  Detail & Related papers  (2024-12-03T07:25:30Z)
• Exploring Gradient Subspaces: Addressing and Overcoming LoRA's Limitations in Federated Fine-Tuning of Large Language Models [19.533062623518674]
  This paper critically analyzes the convergence and performance guarantees of popular FL frameworks utilizing Low-Rank Adaptation (LoRA)<n>We demonstrate that direct weight averaging outperforms LoRA-based strategies, leading to superior performance for fine-tuned models.<n>Our findings show that GaLore along with direct-weight aggregation is a more effective approach, outperforming federated LoRA methods like FlexLoRA and FFA-LoRA across both text and image modalities.
  arXiv  Detail & Related papers  (2024-10-30T15:23:44Z)
• LoRA Done RITE: Robust Invariant Transformation Equilibration for LoRA Optimization [78.93425154518705]
  Low-rank adaption (LoRA) is a widely used parameter-efficient finetuning method for LLM that reduces memory requirements.<n>This paper introduces LoRA-RITE, a novel adaptive matrix preconditioning method for LoRA optimization.
  arXiv  Detail & Related papers  (2024-10-27T22:57:12Z)
• Less is More: Extreme Gradient Boost Rank-1 Adaption for Efficient Finetuning of LLMs [75.11449420928139]
  Fine-tuning Large Language Models (LLMs) has become a crucial technique for adapting pre-trained models to downstream tasks. Low-Rank Adaptation (LoRA) has emerged as a promising solution, but there exists a gap between the practical performance of low-rank adaptations and its theoretical optimum. We propose eXtreme Gradient Boosting LoRA, a novel framework that bridges this gap by leveraging the power of ensemble learning.
  arXiv  Detail & Related papers  (2024-10-25T17:07:13Z)
• MoELoRA: Contrastive Learning Guided Mixture of Experts on Parameter-Efficient Fine-Tuning for Large Language Models [24.17147521556083]
  We introduce a novel PEFT method: MoELoRA. We conduct experiments on 11 tasks in math reasoning and common-sense reasoning benchmarks. MoELoRA achieved an average performance that was 4.2% higher than LoRA, and demonstrated competitive performance compared to the 175B GPT-3.5 on several benchmarks.
  arXiv  Detail & Related papers  (2024-02-20T09:30:48Z)
• DoRA: Weight-Decomposed Low-Rank Adaptation [57.68678247436207]
  We introduce a novel weight decomposition analysis to investigate the inherent differences between FT and LoRA. Aiming to resemble the learning capacity of FT from the findings, we propose Weight-Decomposed Low-Rank Adaptation (DoRA) DoRA decomposes the pre-trained weight into two components, magnitude and direction, for fine-tuning.
  arXiv  Detail & Related papers  (2024-02-14T17:59:34Z)

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.