Copertina

Fast LLM Post-training via Decoupled and Fastest-of-N Speculation

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Cheng, Rongxin, Zhou, Kai, Wei, Xingda, Liu, Siyuan, Han, Mingcong, Ai, Mingjing, Zhou, Yeju, Zhong, Baoquan, Xiao, Wencong, Chen, Rong, Chen, Haibo
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Distributed, Parallel, and Cluster Computing Artificial Intelligence
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

KunServe: Parameter-centric Memory Management for Efficient Memory Overloading Handling in LLM Serving
di: Cheng, Rongxin, et al.
Pubblicazione: (2024)

DecLock: A Case of Decoupled Locking for Disaggregated Memory
di: Zhang, Hanze, et al.
Pubblicazione: (2025)

PhoenixOS: Concurrent OS-level GPU Checkpoint and Restore with Validated Speculation
di: Wei, Xingda, et al.
Pubblicazione: (2024)

DiFache: Efficient and Scalable Caching on Disaggregated Memory using Decentralized Coherence
di: Zhang, Hanze, et al.
Pubblicazione: (2025)

Characterizing the Dilemma of Performance and Index Size in Billion-Scale Vector Search and Breaking It with Second-Tier Memory
di: Cheng, Rongxin, et al.
Pubblicazione: (2024)

ProMoE: Fast MoE-based LLM Serving using Proactive Caching
di: Song, Xiaoniu, et al.
Pubblicazione: (2024)

BLITZSCALE: Fast and Live Large Model Autoscaling with O(1) Host Caching
di: Zhang, Dingyan, et al.
Pubblicazione: (2024)

PipeLLM: Fast and Confidential Large Language Model Services with Speculative Pipelined Encryption
di: Tan, Yifan, et al.
Pubblicazione: (2024)

Ghidorah: Fast LLM Inference on Edge with Speculative Decoding and Hetero-Core Parallelism
di: Wei, Jinhui, et al.
Pubblicazione: (2025)

Accelerating Mixture-of-Experts Inference by Hiding Offloading Latency with Speculative Decoding
di: Wang, Zhibin, et al.
Pubblicazione: (2025)

Towards Lock Modularization for Heterogeneous Environments
di: Zhang, Hanze, et al.
Pubblicazione: (2025)

TensorHub: Scalable and Elastic Weight Transfer for LLM RL Training
di: Ye, Chenhao, et al.
Pubblicazione: (2026)

SPIN: Accelerating Large Language Model Inference with Heterogeneous Speculative Models
di: Chen, Fahao, et al.
Pubblicazione: (2025)

FASER: Fine-Grained Phase Management for Speculative Decoding in Dynamic LLM Serving
di: Chen, Wenyan, et al.
Pubblicazione: (2026)

LMetric: Simple is Better - Multiplication May Be All You Need for LLM Request Scheduling
di: Zhang, Dingyan, et al.
Pubblicazione: (2026)

KVCache Cache in the Wild: Characterizing and Optimizing KVCache Cache at a Large Cloud Provider
di: Wang, Jiahao, et al.
Pubblicazione: (2025)

Speculative Decoding in Decentralized LLM Inference: Turning Communication Latency into Computation Throughput
di: Song, Jingwei, et al.
Pubblicazione: (2025)

Microsecond-scale Dynamic Validation of Idempotency for GPU Kernels
di: Han, Mingcong, et al.
Pubblicazione: (2024)

BurstGPT: A Real-world Workload Dataset to Optimize LLM Serving Systems
di: Wang, Yuxin, et al.
Pubblicazione: (2024)

DistFlow: A Fully Distributed RL Framework for Scalable and Efficient LLM Post-Training
di: Wang, Zhixin, et al.
Pubblicazione: (2025)

Fast State Restoration in LLM Serving with HCache
di: Gao, Shiwei, et al.
Pubblicazione: (2024)

FlexSpec: Frozen Drafts Meet Evolving Targets in Edge-Cloud Collaborative LLM Speculative Decoding
di: Li, Yuchen, et al.
Pubblicazione: (2026)

LLM-CoOpt: A Co-Design and Optimization Framework for Efficient LLM Inference on Heterogeneous Platforms
di: Kong, Jie, et al.
Pubblicazione: (2026)

A Pipelined Collaborative Speculative Decoding Framework for Efficient Edge-Cloud LLM Inference
di: Zhang, Yida, et al.
Pubblicazione: (2026)

Role-Based Fault Tolerance System for LLM RL Post-Training
di: Chen, Zhenqian, et al.
Pubblicazione: (2025)

DualPath: Breaking the Storage Bandwidth Bottleneck in Agentic LLM Inference
di: Wu, Yongtong, et al.
Pubblicazione: (2026)

Fast and Interactive Byzantine Fault-tolerant Web Services via Session-Based Consensus Decoupling
di: Akmal, Ahmad Zaki, et al.
Pubblicazione: (2025)

Compiler Support for Speculation in Decoupled Access/Execute Architectures
di: Szafarczyk, Robert, et al.
Pubblicazione: (2025)

SP-MoE: Speculative Decoding and Prefetching for Accelerating MoE-based Model Inference
di: Chen, Liangkun, et al.
Pubblicazione: (2025)

pBeeGees: A Prudent Approach to Certificate-Decoupled BFT Consensus
di: Yang, Kaiji, et al.
Pubblicazione: (2025)

FlowSpec: Continuous Pipelined Speculative Decoding for Efficient Distributed LLM Inference
di: Liu, Xing, et al.
Pubblicazione: (2025)

SMART: When is it Actually Worth Expanding a Speculative Tree?
di: Wang, Lifu, et al.
Pubblicazione: (2026)

Argus: Token Aware Distributed LLM Inference Optimization
di: Wu, Panlong, et al.
Pubblicazione: (2025)

ReSpec: Towards Optimizing Speculative Decoding in Reinforcement Learning Systems
di: Chen, Qiaoling, et al.
Pubblicazione: (2025)

Fast Iterative Graph Computing with Updated Neighbor States
di: Zhou, Yijie, et al.
Pubblicazione: (2024)

Distributed Speculative Execution for Resilient Cloud Applications
di: Li, Tianyu, et al.
Pubblicazione: (2024)

Chameleon: Taming Dynamic Operator Sequences for Memory-Intensive LLM Training
di: Wang, Zibo, et al.
Pubblicazione: (2025)

FastSet: Parallel Claim Settlement
di: Chen, Xiaohong, et al.
Pubblicazione: (2025)

DualMap: Enabling Both Cache Affinity and Load Balancing for Distributed LLM Serving
di: Yuan, Ying, et al.
Pubblicazione: (2026)

Simulating LLM training workloads for heterogeneous compute and network infrastructure
di: Kumar, Sumit, et al.
Pubblicazione: (2025)