RSQ: Learning from Important Tokens Leads to Better Quantized LLMs

• URL: http://arxiv.org/abs/2503.01820v1
• Date: Mon, 03 Mar 2025 18:46:33 GMT
• Title: RSQ: Learning from Important Tokens Leads to Better Quantized LLMs
• Authors: Yi-Lin Sung, Prateek Yadav, Jialu Li, Jaehong Yoon, Mohit Bansal,
• Abstract summary: Layer-wise quantization is a key technique for efficiently compressing large models without expensive retraining.<n>We propose RSQ (Rotate, Scale, then Quantize), which applies rotations to the model to mitigate outliers.<n>We demonstrate that RSQ consistently outperforms baseline methods across multiple downstream tasks and three model families.
• Score: 65.5558181902098
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Layer-wise quantization is a key technique for efficiently compressing large models without expensive retraining. Previous methods typically quantize the weights of each layer by "uniformly" optimizing the layer reconstruction loss across all output tokens. However, in this paper, we demonstrate that better-quantized models can be obtained by prioritizing learning from important tokens (e.g. which have large attention scores). Building on this finding, we propose RSQ (Rotate, Scale, then Quantize), which (1) applies rotations (orthogonal transformation) to the model to mitigate outliers (those with exceptionally large magnitude), (2) scales the token feature based on its importance, and (3) quantizes the model using the GPTQ framework with the second-order statistics computed by scaled tokens. To compute token importance, we explore both heuristic and dynamic strategies. Based on a thorough analysis of all approaches, we adopt attention concentration, which uses attention scores of each token as its importance, as the best approach. We demonstrate that RSQ consistently outperforms baseline methods across multiple downstream tasks and three model families: LLaMA3, Mistral, and Qwen2.5. Additionally, models quantized with RSQ achieve superior performance on long-context tasks, further highlighting its effectiveness. Lastly, RSQ demonstrates generalizability across various setups, including different model sizes, calibration datasets, bit precisions, and quantization methods.

Related papers

• Enhancing Ultra-Low-Bit Quantization of Large Language Models Through Saliency-Aware Partial Retraining [0.0]
  Post-training quantization reduces model size efficiently at the cost of decreased accuracy. quantization-aware training better preserves accuracy but is resource-intensive. We propose an ultra-low-bit quantization method that builds upon ApiQ and extends its performance without the need for full retraining.
  arXiv  Detail & Related papers  (2025-04-14T19:31:21Z)
• An Efficient Quantum Classifier Based on Hamiltonian Representations [50.467930253994155]
  Quantum machine learning (QML) is a discipline that seeks to transfer the advantages of quantum computing to data-driven tasks. We propose an efficient approach that circumvents the costs associated with data encoding by mapping inputs to a finite set of Pauli strings. We evaluate our approach on text and image classification tasks, against well-established classical and quantum models.
  arXiv  Detail & Related papers  (2025-04-13T11:49:53Z)
• DL-QAT: Weight-Decomposed Low-Rank Quantization-Aware Training for Large Language Models [11.216745641229917]
  Post-training Quantization (PTQ) is a popular technique, but it often faces challenges at low-bit levels. Quantization-aware Training (QAT) can alleviate this problem, but it requires significantly more computational resources. We introduce Weight-Decomposed Low-Rank Quantization-Aware Training (DL-QAT), which merges the advantages of QAT while training only less than 1% of the total parameters.
  arXiv  Detail & Related papers  (2025-04-12T13:57:02Z)
• SKIM: Any-bit Quantization Pushing The Limits of Post-Training Quantization [7.198819240352308]
  Large Language Models (LLMs) exhibit impressive performance across various tasks, but deploying them for inference poses challenges.<n>We propose a new method called SKIM: Scaled K-means clustering wIth Mixed precision.<n>In terms of model perplexity, our method narrows the gap between 3-bit quantized LLaMA models and their full precision counterparts by 16.3% on average.
  arXiv  Detail & Related papers  (2024-12-05T14:19:59Z)
• CLAQ: Pushing the Limits of Low-Bit Post-Training Quantization for LLMs [44.03692512352445]
  Column-Level Adaptive weight Quantization (CLAQ) is a novel and effective framework for Large Language Models (LLMs) quantization. In this paper, we present a novel and effective CLAQ framework by introducing three different types of adaptive strategies for LLM quantization. Experiments on various mainstream open source LLMs including LLaMA-1, LLaMA-2 and Yi demonstrate that our methods achieve the state-of-the-art results across different bit settings.
  arXiv  Detail & Related papers  (2024-05-27T14:49:39Z)
• LLMC: Benchmarking Large Language Model Quantization with a Versatile Compression Toolkit [55.73370804397226]
  Quantization, a key compression technique, can effectively mitigate these demands by compressing and accelerating large language models. We present LLMC, a plug-and-play compression toolkit, to fairly and systematically explore the impact of quantization. Powered by this versatile toolkit, our benchmark covers three key aspects: calibration data, algorithms (three strategies), and data formats.
  arXiv  Detail & Related papers  (2024-05-09T11:49:05Z)
• The Languini Kitchen: Enabling Language Modelling Research at Different Scales of Compute [66.84421705029624]
  We introduce an experimental protocol that enables model comparisons based on equivalent compute, measured in accelerator hours. We pre-process an existing large, diverse, and high-quality dataset of books that surpasses existing academic benchmarks in quality, diversity, and document length. This work also provides two baseline models: a feed-forward model derived from the GPT-2 architecture and a recurrent model in the form of a novel LSTM with ten-fold throughput.
  arXiv  Detail & Related papers  (2023-09-20T10:31:17Z)
• QuantEase: Optimization-based Quantization for Language Models [17.333778751252392]
  This work introduces Quantization (PTQ) of various quantization layers from recent advances of Large Language Models (LLMs) Our CD-based approach features straightforward updates, relying solely on vector operations. We also explore an outlier approach, allowing for retaining significant weights (outoutliers) with complete precision.
  arXiv  Detail & Related papers  (2023-09-05T01:39:09Z)
• 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)
• Boosting Low-Data Instance Segmentation by Unsupervised Pre-training with Saliency Prompt [103.58323875748427]
  This work offers a novel unsupervised pre-training solution for low-data regimes. Inspired by the recent success of the Prompting technique, we introduce a new pre-training method that boosts QEIS models. Experimental results show that our method significantly boosts several QEIS models on three datasets.
  arXiv  Detail & Related papers  (2023-02-02T15:49:03Z)
• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks [133.93803565077337]
  retrieval-augmented generation models combine pre-trained parametric and non-parametric memory for language generation. We show that RAG models generate more specific, diverse and factual language than a state-of-the-art parametric-only seq2seq baseline.
  arXiv  Detail & Related papers  (2020-05-22T21:34:34Z)

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.