论文解读

论文阅读笔记，目前主要关注 LLM / VLM 的量化与 token 压缩；徽章标注中稿会议与年份，没有徽章的为预印本。

21 篇
已中稿 17 篇
Algo21 篇
Sys0 篇
更新于 2026年8月26日

会议 ICLR5 ICML5 CVPR3 NeurIPS2 AAAI1 ACL1 预印本4

分类

2 个

Featured image of post LLM

LLM · 12 篇

Featured image of post VLM

VLM · 9 篇

Algo

算法方向

21 篇

AlgoSOAR: Scale Optimization for Accurate Reconstruction in NVFP4 Quantization

ACL2026AlgoARCQuant: Boosting NVFP4 Quantization with Augmented Residual Channels for LLMs

ICML2026AlgoVEQ: Modality-Adaptive Quantization for MoE Vision-Language Models

NeurIPS2026AlgoBreaking Modality Heterogeneity in Low-Bit Quantization for Large Vision-Language Models

ICML2026AlgoReSpinQuant: Efficient Layer-Wise LLM Quantization via Subspace Residual Rotation Approximation

ICLR2026AlgoSERQ: Saliency-Aware Low-Rank Error Reconstruction For LLM Quantization

ICML2026AlgoOSAQ: Outlier Self-Absorption for Accurate Low-bit LLM Quantization

ICLR2026AlgoSliderQuant: Accurate Post-Training Quantization for LLMs

ICLR2026AlgoTask-related Token Compression in Multi-modal Large Language Models from an Explainability Perspective

AlgoTowards Joint Quantization and Token Pruning of Vision-Language Models

AlgoQAPruner: Quantization-Aware Vision Token Pruning for Multimodal Large Language Models

AlgoVLMQ: Efficient Post-Training Quantization for Large Vision-Language Models via Hessian Augmentation

CVPR2026AlgoFine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients

CVPR2025AlgoMBQ: Modality-Balanced Quantization for Large Vision-Language Models

CVPR2026AlgoMASQuant: Modality-Aware Smoothing Quantization for Multimodal Large Language Models

ICLR2026AlgoTurboQuant: Online Vector Quantization with Near-optimal Distortion Rate

ICLR2023AlgoGPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers

AAAI2024AlgoOWQ: Outlier-Aware Weight Quantization for Efficient Fine-Tuning and Inference of Large Language Models

NeurIPS2022AlgoZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformers

ICML2020AlgoAdaRound: Up or Down? Adaptive Rounding for Post-Training Quantization

ICML2021AlgoAdaQuant: Accurate Post Training Quantization With Small Calibration Sets