Efficiently Aligning Draft Models via Parameter- and Data-Efficient Adaptation

• URL: http://arxiv.org/abs/2603.09527v1
• Date: Tue, 10 Mar 2026 11:35:58 GMT
• Title: Efficiently Aligning Draft Models via Parameter- and Data-Efficient Adaptation
• Abstract summary: Speculative decoding suffers from performance degradation when target models are fine-tuned for specific domains.<n>A naive solution is to retrain draft models for every target model, which is costly and inefficient.<n>We introduce a parameter- and data-efficient framework named Efficient Draft Adaptation, abbreviated as EDA.
• Score: 82.04574413609187
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Speculative decoding accelerates LLM inference but suffers from performance degradation when target models are fine-tuned for specific domains. A naive solution is to retrain draft models for every target model, which is costly and inefficient. To address this, we introduce a parameter- and data-efficient framework named Efficient Draft Adaptation, abbreviated as EDA, for efficiently adapting draft models. EDA introduces three innovations: (1) a decoupled architecture that utilizes shared and private components to model the shared and target-specific output distributions separately, enabling parameter-efficient adaptation by updating only the lightweight private component;(2) a data regeneration strategy that utilizes the fine-tuned target model to regenerate training data, thereby improving the alignment between training and speculative decoding, leading to higher average acceptance length;(3) a sample selection mechanism that prioritizes high-value data for efficient adaptation. Our experiments show that EDA effectively restores speculative performance on fine-tuned models, achieving superior average acceptance lengths with significantly reduced training costs compared to full retraining. Code is available at https://github.com/Lyn-Lucy/Efficient-Draft-Adaptation.

Related papers

• Unsupervised Parameter Efficient Source-free Post-pretraining [52.27955794126508]
  We introduce UpStep, an Unsupervised.<n>Source-free post-pretraining approach to adapt a base model from a source domain to a target domain.<n>We use various general backbone architectures, both supervised and unsupervised, trained on Imagenet as our base model.
  arXiv  Detail & Related papers  (2025-02-28T18:54:51Z)
• Efficient Source-Free Time-Series Adaptation via Parameter Subspace Disentanglement [0.7558576228782637]
  We propose a framework for efficient Source-Free Domain Adaptation (SFDA)<n>Our approach introduces an improved paradigm for source-model preparation and target-side adaptation.<n>We demonstrate that our framework is compatible with various SFDA methods and achieves significant computational efficiency.
  arXiv  Detail & Related papers  (2024-10-03T02:12:03Z)
• Parameter-Efficient Fine-Tuning With Adapters [5.948206235442328]
  This research introduces a novel adaptation method utilizing the UniPELT framework as a base. Our method employs adapters, which enable efficient transfer of pretrained models to new tasks with minimal retraining of the base model parameters.
  arXiv  Detail & Related papers  (2024-05-09T01:40:38Z)
• Federated Learning of Large Language Models with Parameter-Efficient Prompt Tuning and Adaptive Optimization [71.87335804334616]
  Federated learning (FL) is a promising paradigm to enable collaborative model training with decentralized data. The training process of Large Language Models (LLMs) generally incurs the update of significant parameters. This paper proposes an efficient partial prompt tuning approach to improve performance and efficiency simultaneously.
  arXiv  Detail & Related papers  (2023-10-23T16:37:59Z)
• Efficient Adaptation of Large Vision Transformer via Adapter Re-Composing [8.88477151877883]
  High-capacity pre-trained models have revolutionized problem-solving in computer vision. We propose a novel Adapter Re-Composing (ARC) strategy that addresses efficient pre-trained model adaptation. Our approach considers the reusability of adaptation parameters and introduces a parameter-sharing scheme.
  arXiv  Detail & Related papers  (2023-10-10T01:04:15Z)
• Improved Distribution Matching for Dataset Condensation [91.55972945798531]
  We propose a novel dataset condensation method based on distribution matching. Our simple yet effective method outperforms most previous optimization-oriented methods with much fewer computational resources.
  arXiv  Detail & Related papers  (2023-07-19T04:07:33Z)
• LegoNet: A Fast and Exact Unlearning Architecture [59.49058450583149]
  Machine unlearning aims to erase the impact of specific training samples upon deleted requests from a trained model. We present a novel network, namely textitLegoNet, which adopts the framework of fixed encoder + multiple adapters'' We show that LegoNet accomplishes fast and exact unlearning while maintaining acceptable performance, synthetically outperforming unlearning baselines.
  arXiv  Detail & Related papers  (2022-10-28T09:53:05Z)
• Conservative Objective Models for Effective Offline Model-Based Optimization [78.19085445065845]
  Computational design problems arise in a number of settings, from synthetic biology to computer architectures. We propose a method that learns a model of the objective function that lower bounds the actual value of the ground-truth objective on out-of-distribution inputs. COMs are simple to implement and outperform a number of existing methods on a wide range of MBO problems.
  arXiv  Detail & Related papers  (2021-07-14T17:55:28Z)

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.