GAMMA: Global Bit Allocation for Mixed-Precision Models under Arbitrary Budgets

• URL: http://arxiv.org/abs/2605.18475v1
• Date: Mon, 18 May 2026 14:30:58 GMT
• Title: GAMMA: Global Bit Allocation for Mixed-Precision Models under Arbitrary Budgets
• Abstract summary: Mixed-precision quantization improves the budget--accuracy trade-off for large language models.<n>We propose GAMMA, a framework that learns module-wise precision preferences entirely within a post-training pipeline.
• Score: 28.100700446484595
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Mixed-precision quantization improves the budget--accuracy trade-off for large language models (LLMs) by allocating more bits to sensitive modules. However, automating this allocation at LLM scale faces a unique combination of constraints: learnable approaches require quantization-aware training, which is infeasible for billion-parameter models; training-free alternatives rely on static proxy metrics that miss cross-module interactions and must be recomputed per target budget; and search-based methods are expensive without guaranteeing exact budget compliance. We propose GAMMA, a quantizer-agnostic framework that learns module-wise precision preferences entirely within a post-training pipeline. GAMMA optimizes a teacher-forced hidden-state reconstruction objective under an augmented Lagrangian constraint, and projects the learned preferences into exact budget-feasible discrete assignments via integer programming. A key property is score reuse: because the learned preferences encode a stable sensitivity ranking rather than budget-specific weights, a single training run serves arbitrary deployment targets by re-solving only the integer program, reducing per-budget adaptation from hours to a few minutes. Across Llama and Qwen models (8B--32B), GAMMA outperforms both fixed-precision baselines (up to +12.99 Avg.) and search-based mixed-precision methods (up to +7.00 Avg.), and can match fixed 3-bit quality at 2.5-bit average precision, enabling deployment at substantially smaller memory footprints.

Related papers

• QuantWAMs: Calibrating at the Right Granularity for World Action Models [34.540026172139406]
  World Action Models jointly predict future observations and actions, but their iterative denoising and closed-loop execution make deployment costly.<n>We present QuantWAMs, a PTQ framework that aligns quantization decisions with the calibration context defined by model structure, rollout distribution, and task objective.<n>We evaluate QuantWAMs on Fast-WAM and LingBot-VA across RoboTwin 2.0, LIBERO, and real-robot manipulation with an AgiBot G2.
  arXiv  Detail & Related papers  (2026-07-30T15:54:29Z)
• A Feature-Driven Framework for Software Fault Prediction [0.4104921880358479]
  Software fault prediction (SFP) is a critical task in software engineering, enabling early identification of faults in modules to improve software quality and reduce maintenance costs.<n>This research investigates the combined effects of feature selection and parameter tuning on the performance of machine learning (ML) models for SFP.
  arXiv  Detail & Related papers  (2026-05-17T19:16:20Z)
• Model Compression with Exact Budget Constraints via Riemannian Manifolds [39.54576236079211]
  Assigning one of K options to each of N groups under a total cost budget is a recurring problem in efficient AI.<n>We present a new approach by showing that, under softmax relaxation, the budget constraint defines a smooth Riemannian manifold in logit space with unusually simple geometry.<n>Building on these properties, we propose Riemannian Constrained Optimization (RCO), which augments a standard Adam step with tangent projection, binary-search retraction, and momentum transport.
  arXiv  Detail & Related papers  (2026-05-01T13:30:23Z)
• Adaptive Test-Time Compute Allocation for Reasoning LLMs via Constrained Policy Optimization [18.737087162461563]
  Test-time compute scaling has become a powerful lever for improving large language model performance.<n>Yet deploying these techniques under finite inference budgets requires a decision that current systems largely ignore.<n>We formalize this as a constrained optimization problem (maximize expected accuracy subject to an average compute budget) and solve it with a two-stage Solve-then-Learn pipeline.
  arXiv  Detail & Related papers  (2026-04-16T10:39:22Z)
• ScaleBITS: Scalable Bitwidth Search for Hardware-Aligned Mixed-Precision LLMs [14.073708409982705]
  Post-training weight quantization is crucial for reducing the memory and inference cost of large language models (LLMs)<n>We propose ScaleBITS, a mixed-precision quantization framework that enables automated, fine-grained bitwidth allocation under a memory budget.<n> Experiments show that ScaleBITS significantly improves over uniform-precision quantization (up to +36%) and outperforms state-of-the-art sensitivity-aware baselines (up to +13%) in ultra-low-bit regime.
  arXiv  Detail & Related papers  (2026-02-06T18:11:16Z)
• e1: Learning Adaptive Control of Reasoning Effort [88.51897900019485]
  Increasing the thinking budget of AI models can significantly improve accuracy, but not all questions warrant the same amount of reasoning.<n>Users may prefer to allocate different amounts of reasoning effort depending on how they value output quality versus latency and cost.<n>We propose Adaptive Effort Control, a self-adaptive reinforcement learning method that trains models to use a user-specified fraction of tokens.
  arXiv  Detail & Related papers  (2025-10-30T23:12:21Z)
• Achieving PAC Guarantees in Mechanism Design through Multi-Armed Bandits [8.013444110633223]
  We analytically derive a class of optimal solutions to a linear program (LP) for automated mechanism design.<n>These solutions can be expressed using a set of essential variables whose cardinality is exponentially smaller than the total number of variables in the original formulation.<n>We address this by translating the evaluation of this term into a multi-armed bandit (MAB) problem.
  arXiv  Detail & Related papers  (2024-11-30T03:59:36Z)
• HAFLQ: Heterogeneous Adaptive Federated LoRA Fine-tuned LLM with Quantization [55.972018549438964]
  Federated fine-tuning of pre-trained Large Language Models (LLMs) enables task-specific adaptation across diverse datasets while preserving privacy.<n>We propose HAFLQ (Heterogeneous Adaptive Federated Low-Rank Adaptation Fine-tuned LLM with Quantization), a novel framework for efficient and scalable fine-tuning of LLMs in heterogeneous environments.<n> Experimental results on the text classification task demonstrate that HAFLQ reduces memory usage by 31%, lowers communication cost by 49%, improves accuracy by 50%, and achieves faster convergence compared to the baseline method.
  arXiv  Detail & Related papers  (2024-11-10T19:59:54Z)
• Progressive Mixed-Precision Decoding for Efficient LLM Inference [49.05448842542558]
  We introduce Progressive Mixed-Precision Decoding (PMPD) to address the memory-boundedness of decoding.<n>PMPD achieves 1.4$-$12.2$times$ speedup in matrix-vector multiplications over fp16 models.<n>Our approach delivers a throughput gain of 3.8$-$8.0$times$ over fp16 models and up to 1.54$times$ over uniform quantization approaches.
  arXiv  Detail & Related papers  (2024-10-17T11:46:33Z)
• Decoding-Time Language Model Alignment with Multiple Objectives [116.42095026960598]
  Existing methods primarily focus on optimizing LMs for a single reward function, limiting their adaptability to varied objectives. Here, we propose $textbfmulti-objective decoding (MOD)$, a decoding-time algorithm that outputs the next token from a linear combination of predictions. We show why existing approaches can be sub-optimal even in natural settings and obtain optimality guarantees for our method.
  arXiv  Detail & Related papers  (2024-06-27T02:46:30Z)
• LoRA-Ensemble: Efficient Uncertainty Modelling for Self-Attention Networks [52.46420522934253]
  We introduce LoRA-Ensemble, a parameter-efficient ensembling method for self-attention networks.<n>The method not only outperforms state-of-the-art implicit techniques like BatchEnsemble, but even matches or exceeds the accuracy of an Explicit Ensemble.
  arXiv  Detail & Related papers  (2024-05-23T11:10:32Z)
• MILO: Model-Agnostic Subset Selection Framework for Efficient Model Training and Tuning [68.12870241637636]
  We propose MILO, a model-agnostic subset selection framework that decouples the subset selection from model training. Our empirical results indicate that MILO can train models $3times - 10 times$ faster and tune hyperparameters $20times - 75 times$ faster than full-dataset training or tuning without performance.
  arXiv  Detail & Related papers  (2023-01-30T20:59:30Z)

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.