Quantization Beyond Uniform Bit Allocation

• URL: http://arxiv.org/abs/2608.19388v1
• Date: Wed, 19 Aug 2026 19:06:43 GMT
• Title: Quantization Beyond Uniform Bit Allocation
• Abstract summary: We investigate whether a variable bit allocation scheme can improve quantization quality under a fixed memory budget.<n>We propose a simple variable bit allocation framework that partitions an embedding into contiguous buckets and allocates storage non-uniformly across them.<n>Our results suggest a new direction for structure-aware compression and indexing techniques for large-scale retrieval systems.
• Score: 20.940721219907804
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Quantization is a fundamental technique to handle the growing sizes of embeddings generated by modern models. Existing quantization schemes are largely embedding agnostic and allocate bits uniformly across dimensions. However, recent models produce embeddings with significant geometric structure. In this work, we investigate whether a variable bit allocation scheme can improve quantization quality under a fixed memory budget. We propose a simple variable bit allocation framework that partitions an embedding into contiguous buckets and allocates storage non-uniformly across them. Using a greedy allocation strategy, we instantiate this framework for both Product Quantization (PQ) and Scalar Quantization (SQ). We perform a series of experiments on embeddings known to have the Matryoshka property (MRL), and consistently observe that non-uniform allocations outperform uniform baselines at identical storage budgets. The largest improvements occur in the low-bit regime, where uniform allocation is particularly inefficient for MRL embeddings. At the same compression rates, variable allocation improves recall by up to 8\% for PQ and up to 18\% for SQ. Our results suggest a new direction for structure-aware compression and indexing techniques for large-scale retrieval systems.

Related papers

• Variable-Width Transformers [57.07167443557886]
  We investigate nonuniform capacity allocation across network depth by proposing a $times$-shaped > former architecture.<n>Our design maintains wider early and late layers while narrowing the middle layers, utilizing a parameter-free residual resizing mechanism.<n>By reducing the average layer width, this architecture also requires fewer overall FLOPs and smaller KV cache memory and I/O cost.
  arXiv  Detail & Related papers  (2026-06-16T17:59:03Z)
• BPDQ: Bit-Plane Decomposition Quantization on a Variable Grid for Large Language Models [56.504879072674015]
  We propose Bit-Plane Decomposition Quantization (BPDQ), which constructs a variable quantization grid via bit-planes and scalar coefficients.<n>BPDQ enables serving Qwen2.5-72B on a single GTX 3090 with 83.85% GSM8K accuracy (vs. 90.83% at 16-bit)
  arXiv  Detail & Related papers  (2026-02-04T02:54:37Z)
• A Clustering-Based Variable Ordering Framework for Relaxed Decision Diagrams for Maximum Weighted Independent Set Problem [4.312746668772342]
  This work introduces a novel clustering-based framework for variable ordering.<n>Instead of applying dynamic orderings to the full set of unfixed variables, we first primal variables into clusters.<n>We then leverage this structural decomposition to guide the ordering process, significantly reducing the partition's search space.
  arXiv  Detail & Related papers  (2025-12-17T08:49:38Z)
• Learning Grouped Lattice Vector Quantizers for Low-Bit LLM Compression [57.54335545892155]
  We introduce a Grouped Lattice Vector Quantization (GLVQ) framework that assigns each group of weights a customized lattice codebook.<n>Our approach achieves a better trade-off between model size and accuracy compared to existing post-training quantization baselines.
  arXiv  Detail & Related papers  (2025-10-23T20:19:48Z)
• BitStack: Any-Size Compression of Large Language Models in Variable Memory Environments [53.71158537264695]
  Large language models (LLMs) have revolutionized numerous applications, yet their deployment remains challenged by memory constraints on local devices.<n>We introduce textbfBitStack, a novel, training-free weight compression approach that enables megabyte-level trade-offs between memory usage and model performance.
  arXiv  Detail & Related papers  (2024-10-31T13:26:11Z)
• SliM-LLM: Salience-Driven Mixed-Precision Quantization for Large Language Models [63.118592279833656]
  Post-training quantization (PTQ) is an effective technique for compressing large language models (LLMs)<n>We propose SliM-LLM, a salience-driven mixed-precision quantization framework that allocates bit-widths at the group-wise.<n> Experiments show that SliM-LLM achieves superior performance across various LLMs at low bit-widths.
  arXiv  Detail & Related papers  (2024-05-23T16:21:48Z)
• SqueezeLLM: Dense-and-Sparse Quantization [80.32162537942138]
  Main bottleneck for generative inference with LLMs is memory bandwidth, rather than compute, for single batch inference. We introduce SqueezeLLM, a post-training quantization framework that enables lossless compression to ultra-low precisions of up to 3-bit. Our framework incorporates two novel ideas: (i) sensitivity-based non-uniform quantization, which searches for the optimal bit precision assignment based on second-order information; and (ii) the Dense-and-Sparse decomposition that stores outliers and sensitive weight values in an efficient sparse format.
  arXiv  Detail & Related papers  (2023-06-13T08:57:54Z)
• SDQ: Stochastic Differentiable Quantization with Mixed Precision [46.232003346732064]
  We present a novel Differentiable Quantization (SDQ) method that can automatically learn the MPQ strategy. After the optimal MPQ strategy is acquired, we train our network with entropy-aware bin regularization and knowledge distillation. SDQ outperforms all state-of-the-art mixed datasets or single precision quantization with a lower bitwidth.
  arXiv  Detail & Related papers  (2022-06-09T12:38:18Z)

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.