89 citations · 194 across the 58 of their papers we have counts for
62 papers
PDD: Unleashing Economical and Flexible Heterogeneous LLM Inference via Cross-Datacenter Prefill-Decode Disaggregation
Yida Wang, Xiuhong Li, Jianping Ma +9
Interconnecting geographically dispersed clusters over wide-area Ethernet provides a scalable and cost-effective alternative to dedicated intra-datacenter heterogeneous clusters fo…
Claw-SWE-Bench: A Benchmark for Evaluating OpenClaw-style Agent Harnesses on Coding Tasks
Mengyu Zheng, Kai Han, Boxun Li +13
The software engineering capabilities of general-purpose agent harnesses remain underexplored, and existing benchmarks offer limited support for comparing these harnesses under con…
Infini Memory: Maintainable Topic Documents for Long-Term LLM Agent Memory
Suozhao Ji, Baodong Wu, Zehao Wang +8
Long-term LLM agents need persistent memory that can track changing facts and provide relevant evidence across sessions. Existing memory systems often store observations as isolate…
Elastic-dLLM: Position Preserving Context Compression and Augmentation of Diffusion LLMs
Junyi Wu, Tianchen Zhao, Shaoqiu Zhang +3
Unlike autoregressive models, which generate one token at a time, dLLMs denoise a chunk of [MASK] tokens jointly and sample one or more tokens per step; despite enabling parallel d…
DynaTrain: Fast Online Parallelism Switching for Elastic LLM Training
Yuanqing Wang, Yuchen Zhang, Hao Lin +9
Modern large language model (LLM) training is inherently dynamic: resource fluctuations, RLHF phase shifts, and cluster elasticity continually reshape the optimal parallelism layou…
FUSCO: High-Performance Distributed Data Shuffling via Transformation-Communication Fusion
Zhuoran Zhu, Chunyang Zhu, Hao Lin +9
Large-scale Mixture-of-Experts (MoE) models rely on \emph{expert parallelism} for efficient training and inference, which splits experts across devices and necessitates distributed…
SJD++: Improved Speculative Jacobi Decoding for Training-free Acceleration of Discrete Auto-regressive Text-to-Image Generation
Yao Teng, Zhihuan Jiang, Han Shi +6
Large autoregressive models can generate high-quality, high-resolution images but suffer from slow generation speed, because these models require hundreds to thousands of sequentia…
Reducing Latency of LLM Search Agent via Speculation-based Algorithm-System Co-Design
Zixiao Huang, Wen Zeng, Tianyu Fu +10
LLM-based search agents achieve strong performance but suffer from severe latency, as each step requires serialized LLM reasoning followed by action of tool execution. We revisit t…
BitSnap: Checkpoint Sparsification and Quantization in LLM Training
Yanxin Peng, Qingping Li, Baodong Wu +4
As large language models (LLMs) continue to grow in size and complexity, efficient checkpoint saving\&loading has become crucial for managing storage, memory usage, and fault toler…
Think-at-Hard: Dynamic Looped Transformers for Improved Reasoning
Tianyu Fu, Yichen You, Zekai Chen +3
Improving the reasoning abilities of Large Language Models (LLMs), especially under parameter constraints, is crucial for real-world applications. Looped transformers address this…
Distilled Decoding 2: One-step Sampling of Image Auto-regressive Models with Conditional Score Distillation
Enshu Liu, Qian Chen, Xuefei Ning +4
Image Auto-regressive (AR) models have emerged as a powerful paradigm of visual generative models. Despite their promising performance, they suffer from slow generation speed due t…
ReXMoE: Reusing Experts with Minimal Overhead in Mixture-of-Experts
Zheyue Tan, Zhiyuan Li, Tao Yuan +13
Mixture-of-Experts (MoE) architectures have emerged as a promising approach to scale Large Language Models (LLMs). MoE boosts the efficiency by activating a subset of experts per t…
TASP: Topology-aware Sequence Parallelism
Yida Wang, Ke Hong, Xiuhong Li +4
Long-context large language models (LLMs) face constraints due to the quadratic complexity of the self-attention mechanism. The mainstream sequence parallelism (SP) method, Ring At…
Cache-to-Cache: Direct Semantic Communication Between Large Language Models
Tianyu Fu, Zihan Min, Hanling Zhang +4
Multi-LLM systems harness the complementary strengths of diverse Large Language Models, achieving performance and efficiency gains that are not attainable by a single model. In exi…
RLinf: Flexible and Efficient Large-scale Reinforcement Learning via Macro-to-Micro Flow Transformation
Chao Yu, Yuanqing Wang, Zhen Guo +26
Reinforcement learning (RL) has demonstrated immense potential in advancing artificial general intelligence, agentic intelligence, and embodied intelligence. However, the inherent…
VocabTailor: Dynamic Vocabulary Selection for Downstream Tasks in Small Language Models
Hanling Zhang, Yayu Zhou, Tongcheng Fang +4
Small Language Models (SLMs) provide computational advantages in resource-constrained environments, yet memory limitations remain a critical bottleneck for edge device deployment.…
STAlloc: Enhancing Memory Efficiency in Large-Scale Model Training with Spatio-Temporal Planning
Zixiao Huang, Junhao Hu, Hao Lin +9
The rapid scaling of large language models (LLMs) has significantly increased GPU memory pressure, which is further aggravated by training optimization techniques such as virtual p…
R2R: Efficiently Navigating Divergent Reasoning Paths with Small-Large Model Token Routing
Tianyu Fu, Yi Ge, Yichen You +6
Large Language Models (LLMs) achieve impressive reasoning capabilities at the cost of substantial inference overhead, posing substantial deployment challenges. Although distilled S…
PM-KVQ: Progressive Mixed-precision KV Cache Quantization for Long-CoT LLMs
Tengxuan Liu, Shiyao Li, Jiayi Yang +7
Recently, significant progress has been made in developing reasoning-capable Large Language Models (LLMs) through long Chain-of-Thought (CoT) techniques. However, this long-CoT rea…
semi-PD: Towards Efficient LLM Serving via Phase-Wise Disaggregated Computation and Unified Storage
Ke Hong, Lufang Chen, Zhong Wang +9
Existing large language model (LLM) serving systems fall into two categories: 1) a unified system where prefill phase and decode phase are co-located on the same GPU, sharing the u…
VGDFR: Diffusion-based Video Generation with Dynamic Latent Frame Rate
Zhihang Yuan, Rui Xie, Yuzhang Shang +5
Diffusion Transformer(DiT)-based generation models have achieved remarkable success in video generation. However, their inherent computational demands pose significant efficiency c…
DLFR-VAE: Dynamic Latent Frame Rate VAE for Video Generation
Zhihang Yuan, Siyuan Wang, Rui Xie +6
In this paper, we propose the Dynamic Latent Frame Rate VAE (DLFR-VAE), a training-free paradigm that can make use of adaptive temporal compression in latent space. While existing…
Efficient and Adaptable Overlapping for Computation and Communication via Signaling and Reordering
Ke Hong, Xiuhong Li, Minxu Liu +9
Generative models have achieved remarkable success across various applications, driving the demand for multi-GPU computing. Inter-GPU communication becomes a bottleneck in multi-GP…
DiTFastAttnV2: Head-wise Attention Compression for Multi-Modality Diffusion Transformers
Hanling Zhang, Rundong Su, Zhihang Yuan +5
Text-to-image generation models, especially Multimodal Diffusion Transformers (MMDiT), have shown remarkable progress in generating high-quality images. However, these models often…