Fitting Is Not Enough: Smoothness in Extremely Quantized LLMs

• URL: http://arxiv.org/abs/2605.08894v2
• Date: Fri, 15 May 2026 02:06:51 GMT
• Title: Fitting Is Not Enough: Smoothness in Extremely Quantized LLMs
• Abstract summary: Large language models (LLMs) achieve strong performance but incur high deployment costs.<n>Existing quantization algorithms mainly focus on improving the numerical accuracy of forward computation.<n>We show that extremely quantized LLMs suffer from systematic degradation beyond numerical precision loss.
• Score: 46.88109820641649
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) achieve strong performance but incur high deployment costs, motivating extremely low-bit but lossy quantization. Existing quantization algorithms mainly focus on improving the numerical accuracy of forward computation to eliminate performance degradation. In this paper, we show that extremely quantized LLMs suffer from systematic smoothness degradation beyond numerical precision loss. Through a smoothness proxy, we observe that such degradation becomes increasingly severe as the quantization bit-width decreases. Furthermore, based on sequence neighborhood modeling, we find that quantized models exhibit a rapid reduction of effective token candidates within the prediction neighborhood, which directly leads to a sparser decoding tree and degraded generation quality. To validate it, we introduce a simple smoothness-preserving principle in both post-training quantization and quantization-aware training, and demonstrate that preserving smoothness brings additional gains beyond numerical accuracy. The core goal of this paper is to highlight smoothness preservation as an important design consideration for future extreme quantization methods. Code is available at https://github.com/xuyuzhuang11/FINE.

Related papers

• CodeQuant: Unified Clustering and Quantization for Enhanced Outlier Smoothing in Low-Precision Mixture-of-Experts [11.005678744572888]
  Outliers have emerged as a fundamental bottleneck in preserving accuracy for low-precision large models.<n>We introduce CodeQuant, a unified quantization-and-clustering scheme that contains smoothing activation outliers via learnable rotation and absorbing weight outliers into fine-tuned cluster centroids for MoE.<n>CodeQuant achieves up to $4.15times$ speedup while delivering significantly higher accuracy than state-of-the-art quantization approaches across diverse MoE models.
  arXiv  Detail & Related papers  (2026-04-12T07:06:16Z)
• End-to-End On-Device Quantization-Aware Training for LLMs at Inference Cost [53.25965863436039]
  Quantization-aware training (QAT) provides a more principled solution, but its reliance on backpropagation incurs prohibitive memory costs.<n>We propose ZeroQAT, a zeroth-order optimization-based QAT framework that supports both weight and activation quantization.<n>Experiments show that ZeroQAT consistently outperforms representative PTQ and QAT baselines while requiring significantly less memory.
  arXiv  Detail & Related papers  (2025-08-21T01:18:27Z)
• MPQ-DMv2: Flexible Residual Mixed Precision Quantization for Low-Bit Diffusion Models with Temporal Distillation [74.34220141721231]
  We present MPQ-DMv2, an improved textbfMixed textbfPrecision textbfQuantization framework for extremely low-bit textbfDiffusion textbfModels.
  arXiv  Detail & Related papers  (2025-07-06T08:16:50Z)
• FIMA-Q: Post-Training Quantization for Vision Transformers by Fisher Information Matrix Approximation [55.12070409045766]
  Post-training quantization (PTQ) has stood out as a cost-effective and promising model compression paradigm in recent years.<n>Current PTQ methods for Vision Transformers (ViTs) still suffer from significant accuracy degradation, especially under low-bit quantization.
  arXiv  Detail & Related papers  (2025-06-13T07:57:38Z)
• Understanding the Difficulty of Low-Precision Post-Training Quantization for LLMs [4.5529796609245805]
  Large language models of high parameter counts are computationally expensive, yet can be made much more efficient by compressing their weights to very low numerical precision.<n>We show that, under the same data constraint, the former approach nearly always fared worse than the latter, a phenomenon particularly prominent when the numerical precision is very low.
  arXiv  Detail & Related papers  (2024-10-18T16:16:52Z)
• Compensate Quantization Errors+: Quantized Models Are Inquisitive Learners [51.32182730502002]
  We introduce Singular-value Diagonal Expansion to refine weight distributions to achieve better quantization alignment.<n>Our plug-and-play weight-quantization methods demonstrate substantial performance improvements over state-of-the-art approaches.
  arXiv  Detail & Related papers  (2024-07-22T09:45:16Z)
• WKVQuant: Quantizing Weight and Key/Value Cache for Large Language Models Gains More [55.0856305773081]
  Large Language Models (LLMs) face significant deployment challenges due to their substantial memory requirements and the computational demands of auto-regressive text generation process. This paper addresses these challenges by focusing on the quantization of LLMs, a technique that reduces memory consumption by converting model parameters and activations into low-bit integers.
  arXiv  Detail & Related papers  (2024-02-19T11:33:21Z)
• ApiQ: Finetuning of 2-Bit Quantized Large Language Model [12.328293460903911]
  ApiQ is designed to restore the lost information from quantization by concurrently initializing the LoRA components and quantizing the weights of LLMs. It consistently achieves superior finetuning results across various bit-widths.
  arXiv  Detail & Related papers  (2024-02-07T09:36:54Z)
• Zero-Shot Sharpness-Aware Quantization for Pre-trained Language Models [88.80146574509195]
  Quantization is a promising approach for reducing memory overhead and accelerating inference. We propose a novel-aware quantization (ZSAQ) framework for the zero-shot quantization of various PLMs.
  arXiv  Detail & Related papers  (2023-10-20T07:09:56Z)
• VecQ: Minimal Loss DNN Model Compression With Vectorized Weight Quantization [19.66522714831141]
  We develop a new quantization solution called VecQ, which can guarantee minimal direct quantization loss and better model accuracy. In addition, in order to up the proposed quantization process during training, we accelerate the quantization process with a parameterized estimation and probability-based calculation.
  arXiv  Detail & Related papers  (2020-05-18T07:38:44Z)

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.