Image de couverture de livre

Scaling Up On-Device LLMs via Active-Weight Swapping Between DRAM and Flash

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Jia, Fucheng, Wu, Zewen, Jiang, Shiqi, Jiang, Huiqiang, Zhang, Qianxi, Yang, Yuqing, Liu, Yunxin, Ren, Ju, Zhang, Deyu, Cao, Ting
Format:	Preprint
Publié:	2025
Sujets:	Machine Learning
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Empowering In-Browser Deep Learning Inference on Edge Devices with Just-in-Time Kernel Optimizations
par: Jia, Fucheng, et autres
Publié: (2023)

Scaling LLM Test-Time Compute with Mobile NPU on Smartphones
par: Hao, Zixu, et autres
Publié: (2025)

AdaNav: Adaptive Reasoning with Uncertainty for Vision-Language Navigation
par: Ding, Xin, et autres
Publié: (2025)

Em-Garde: A Propose-Match Framework for Proactive Streaming Video Understanding
par: Zheng, Yikai, et autres
Publié: (2026)

Making Every Frame Matter: Continuous Activity Recognition in Streaming Video via Adaptive Video Context Modeling
par: Wu, Hao, et autres
Publié: (2024)

KVNAND: Efficient On-Device Large Language Model Inference Using DRAM-Free In-Flash Computing
par: Deng, Lishuo, et autres
Publié: (2025)

A First Look At Efficient And Secure On-Device LLM Inference Against KV Leakage
par: Yang, Huan, et autres
Publié: (2024)

MemCompiler: Compile, Don't Inject -- State-Conditioned Memory for Embodied Agents
par: Ding, Xin, et autres
Publié: (2026)

Sculptor: Empowering LLMs with Cognitive Agency via Active Context Management
par: Li, Mo, et autres
Publié: (2025)

EmbodiSkill: Skill-Aware Reflection for Self-Evolving Embodied Agents
par: Ju, Ruofei, et autres
Publié: (2026)

Accelerating Prefilling via Decoding-time Contribution Sparsity
par: He, Zhiyuan, et autres
Publié: (2025)

MTraining: Distributed Dynamic Sparse Attention for Efficient Ultra-Long Context Training
par: Li, Wenxuan, et autres
Publié: (2025)

MobiFuse: A High-Precision On-device Depth Perception System with Multi-Data Fusion
par: Zhang, Jinrui, et autres
Publié: (2024)

SwapNet: Efficient Swapping for DNN Inference on Edge AI Devices Beyond the Memory Budget
par: Wang, Kun, et autres
Publié: (2024)

LongLLMLingua: Accelerating and Enhancing LLMs in Long Context Scenarios via Prompt Compression
par: Jiang, Huiqiang, et autres
Publié: (2023)

Uni-MoE: Scaling Unified Multimodal LLMs with Mixture of Experts
par: Li, Yunxin, et autres
Publié: (2024)

SortedRL: Accelerating RL Training for LLMs through Online Length-Aware Scheduling
par: Zhang, Yiqi, et autres
Publié: (2026)

FP-Rowhammer: DRAM-Based Device Fingerprinting
par: Venugopalan, Hari, et autres
Publié: (2023)

JoyAI-LLM Flash: Advancing Mid-Scale LLMs with Token Efficiency
par: Cai, Aichen, et autres
Publié: (2026)

LeanK: Learnable K Cache Channel Pruning for Efficient Decoding
par: Zhang, Yike, et autres
Publié: (2025)

Revisiting DRAM Read Disturbance: Identifying Inconsistencies Between Experimental Characterization and Device-Level Studies
par: Luo, Haocong, et autres
Publié: (2025)

Settling Weighted Token Swapping up to Algorithmic Barriers
par: Wein, Nicole, et autres
Publié: (2025)

Weisfeiler‐Lehman Kernel Augmented Product Representation for Queries on Large‐Scale BIM Scenes
par: Huiqiang Hu, et autres
Publié: (2025)

FlashRecovery: Fast and Low-Cost Recovery from Failures for Large-Scale Training of LLMs
par: Zhang, Haijun, et autres
Publié: (2025)

Distinguishing Translations by Human, NMT, and ChatGPT: A Linguistic and Statistical Approach
par: Jiang, Zhaokun, et autres
Publié: (2023)

Convergences and Divergences between Automatic Assessment and Human Evaluation: Insights from Comparing ChatGPT-Generated Translation and Neural Machine Translation
par: Jiang, Zhaokun, et autres
Publié: (2024)

RetrievalAttention: Accelerating Long-Context LLM Inference via Vector Retrieval
par: Liu, Di, et autres
Publié: (2024)

Uncovering the Role of Initial Saliency in U-Shaped Attention Bias: Scaling Initial Token Weight for Enhanced Long-Text Processing
par: Qiang, Zewen, et autres
Publié: (2025)

Hardware Co-Design Scaling Laws via Roofline Modelling for On-Device LLMs
par: Sun, Luoyang, et autres
Publié: (2026)

Position Engineering: Boosting Large Language Models through Positional Information Manipulation
par: He, Zhiyuan, et autres
Publié: (2024)

An Analytic Model to Determine the Interstitial-Solute Energetics and Underlying Mechanism in Refractory High-Entropy Alloys
par: Zhu, Qianxi, et autres
Publié: (2025)

Vec-LUT: Vector Table Lookup for Parallel Ultra-Low-Bit LLM Inference on Edge Devices
par: Li, Xiangyu, et autres
Publié: (2025)

Zoomer: Adaptive Image Focus Optimization for Black-box MLLM
par: Qian, Jiaxu, et autres
Publié: (2025)

ProRe: A Proactive Reward System for GUI Agents via Reasoner-Actor Collaboration
par: Dai, Gaole, et autres
Publié: (2025)

Advancing Mobile GUI Agents: A Verifier-Driven Approach to Practical Deployment
par: Dai, Gaole, et autres
Publié: (2025)

AVA: Towards Agentic Video Analytics with Vision Language Models
par: Yan, Yuxuan, et autres
Publié: (2025)

PUDTune: Multi-Level Charging for High-Precision Calibration in Processing-Using-DRAM
par: Kubo, Tatsuya, et autres
Publié: (2025)

SecurityLingua: Efficient Defense of LLM Jailbreak Attacks via Security-Aware Prompt Compression
par: Li, Yucheng, et autres
Publié: (2025)

Mitigate Position Bias in Large Language Models via Scaling a Single Dimension
par: Yu, Yijiong, et autres
Publié: (2024)

LLMs Meet Long Video: Advancing Long Video Question Answering with An Interactive Visual Adapter in LLMs
par: Li, Yunxin, et autres
Publié: (2024)