KurTail : Kurtosis-based LLM Quantization

• URL: http://arxiv.org/abs/2503.01483v1
• Date: Mon, 03 Mar 2025 12:43:06 GMT
• Title: KurTail : Kurtosis-based LLM Quantization
• Authors: Mohammad Sadegh Akhondzadeh, Aleksandar Bojchevski, Evangelos Eleftheriou, Martino Dazzi,
• Abstract summary: KurTail is a new post-training quantization scheme that mitigates outliers in the activations of large language models.<n>It offers a 13.3% boost in MMLU accuracy and a 15.5% drop in Wiki perplexity compared to QuaRot.<n>It also outperforms SpinQuant with a 2.6% MMLU gain and reduces perplexity by 2.9%, all while reducing the training cost.
• Score: 51.24081396305435
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: One of the challenges of quantizing a large language model (LLM) is the presence of outliers. Outliers often make uniform quantization schemes less effective, particularly in extreme cases such as 4-bit quantization. We introduce KurTail, a new post-training quantization (PTQ) scheme that leverages Kurtosis-based rotation to mitigate outliers in the activations of LLMs. Our method optimizes Kurtosis as a measure of tailedness. This approach enables the quantization of weights, activations, and the KV cache in 4 bits. We utilize layer-wise optimization, ensuring memory efficiency. KurTail outperforms existing quantization methods, offering a 13.3\% boost in MMLU accuracy and a 15.5\% drop in Wiki perplexity compared to QuaRot. It also outperforms SpinQuant with a 2.6\% MMLU gain and reduces perplexity by 2.9\%, all while reducing the training cost. For comparison, learning the rotation using SpinQuant for Llama3-70B requires at least four NVIDIA H100 80GB GPUs, whereas our method requires only a single GPU, making it a more accessible solution for consumer GPU.

Related papers

• SVDQuant: Absorbing Outliers by Low-Rank Components for 4-Bit Diffusion Models [58.5019443418822]
  Diffusion models can generate high-quality images, but as they scale, rising memory demands and higher latency pose deployment challenges.<n>We propose SVDQuant, a new 4-bit quantization paradigm to overcome this limitation.<n>We reduce the memory usage for the 12B FLUX.1 models by 3.5$times$, achieving 3.0$times$ speedup over the 4-bit weight-only quantization (W4A16) baseline.
  arXiv  Detail & Related papers  (2024-11-07T18:59:58Z)
• FlatQuant: Flatness Matters for LLM Quantization [58.28221892035609]
  We propose FlatQuant, a new post-training quantization approach to enhance flatness of weights and activations. Our approach identifies optimal affine transformations tailored to each linear layer, calibrated in hours via a lightweight objective runtime. For inference latency, FlatQuant reduces the slowdown induced by pre-quantization transformation from 0.26x of QuaRot to merely $textbf0.07x$, bringing up to $textbf2.3x$ speedup for prefill and $textbf1.7x$ speedup for decoding.
  arXiv  Detail & Related papers  (2024-10-12T08:10:28Z)
• SpinQuant: LLM quantization with learned rotations [49.07335692298487]
  Post-training quantization (PTQ) techniques applied to weights, activations, and the KV cache greatly reduce memory usage, latency, and power consumption of Large Language Models (LLMs)<n>We identify a collection of applicable rotation parameterizations that lead to identical outputs in full-precision Transformer architectures while enhancing quantization accuracy.<n>We propose SpinQuant, a novel approach that incorporates learned rotation matrices for optimal quantized network accuracy.
  arXiv  Detail & Related papers  (2024-05-26T02:15:49Z)
• QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving [52.31791050376249]
  Quantization can accelerate large language model (LLM) inference. We introduce QoQ, a W4A8KV4 quantization algorithm with 4-bit weight, 8-bit activation, and 4-bit KV cache. QServe improves the maximum achievable serving of Llama-3-8B by 1.2x on A100, 1.4x on L40S; and Qwen-721.5B by 2.4x on A100, 3.5x on L40S.
  arXiv  Detail & Related papers  (2024-05-07T17:59:30Z)
• KVQuant: Towards 10 Million Context Length LLM Inference with KV Cache Quantization [67.74400574357472]
  LLMs are seeing growing use for applications which require large context windows, and with these large context windows KV cache activations surface as the dominant contributor to memory consumption during inference. Quantization is a promising approach for compressing KV cache activations; however, existing solutions fail to represent activations accurately in sub-4-bit precision. Our work, KVQuant, facilitates low precision KV cache quantization by incorporating several novel methods.
  arXiv  Detail & Related papers  (2024-01-31T18:58:14Z)
• SmoothQuant+: Accurate and Efficient 4-bit Post-Training WeightQuantization for LLM [13.035063417593534]
  Large language models (LLMs) have shown remarkable capabilities in various tasks. Currently, 4-bit post-training quantization (PTQ) has achieved some success in LLMs. We propose SmoothQuant+, an accurate and efficient 4-bit weight-only PTQ.
  arXiv  Detail & Related papers  (2023-12-06T11:10:55Z)
• Fast and Efficient 2-bit LLM Inference on GPU: 2/4/16-bit in a Weight Matrix with Asynchronous Dequantization [14.201092042777299]
  Large language models (LLMs) have demonstrated impressive abilities in various domains while the inference cost is expensive. Applying 2-bit single-precision weight quantization brings >3% accuracy loss. We propose mixed-precision quantization for each weight matrix and asynchronous dequantization during inference.
  arXiv  Detail & Related papers  (2023-11-28T02:44:59Z)

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.