492 citations · 1.5k across the 65 of their papers we have counts for
25 papers · 1 filter
HashAttention: Semantic Sparsity for Faster Inference
Aditya Desai, Shuo Yang, Alejandro Cuadron +3
Leveraging long contexts is crucial for advanced AI systems, but attention computation poses a scalability challenge. While scaled dot-product attention (SDPA) exhibits token spars…
BlendServe: Optimizing Offline Inference for Auto-regressive Large Models with Resource-aware Batching
Yilong Zhao, Shuo Yang, Kan Zhu +5
Offline batch inference, which leverages the flexibility of request batching to achieve higher throughput and lower costs, is becoming more popular for latency-insensitive applicat…
Efficient LLM Scheduling by Learning to Rank
Yichao Fu, Siqi Zhu, Runlong Su +3
In Large Language Model (LLM) inference, the output length of an LLM request is typically regarded as not known a priori. Consequently, most LLM serving systems employ a simple Fir…
Post-Training Sparse Attention with Double Sparsity
Shuo Yang, Ying Sheng, Joseph E. Gonzalez +2
The inference process for large language models is slow and memory-intensive, with one of the most critical bottlenecks being excessive Key-Value (KV) cache accesses. This paper in…
RouteLLM: Learning to Route LLMs with Preference Data
Isaac Ong, Amjad Almahairi, Vincent Wu +5
Large language models (LLMs) exhibit impressive capabilities across a wide range of tasks, yet the choice of which model to use often involves a trade-off between performance and c…
From Crowdsourced Data to High-Quality Benchmarks: Arena-Hard and BenchBuilder Pipeline
Tianle Li, Wei-Lin Chiang, Evan Frick +5
The rapid evolution of Large Language Models (LLMs) has outpaced the development of model evaluation, highlighting the need for continuous curation of new, challenging benchmarks.…
Break the Sequential Dependency of LLM Inference Using Lookahead Decoding
Yichao Fu, Peter Bailis, Ion Stoica +1
Autoregressive decoding of large language models (LLMs) is memory bandwidth bounded, resulting in high latency and significant wastes of the parallel processing power of modern acc…
LLM-Assisted Code Cleaning For Training Accurate Code Generators
Naman Jain, Tianjun Zhang, Wei-Lin Chiang +3
Natural language to code generation is an important application area of LLMs and has received wide attention from the community. The majority of relevant studies have exclusively c…
S-LoRA: Serving Thousands of Concurrent LoRA Adapters
Ying Sheng, Shiyi Cao, Dacheng Li +9
The "pretrain-then-finetune" paradigm is commonly adopted in the deployment of large language models. Low-Rank Adaptation (LoRA), a parameter-efficient fine-tuning method, is often…
DISTFLASHATTN: Distributed Memory-efficient Attention for Long-context LLMs Training
Dacheng Li, Rulin Shao, Anze Xie +5
FlashAttention (Dao, 2023) effectively reduces the quadratic peak memory usage to linear in training transformer-based large language models (LLMs) on a single GPU. In this paper,…
Efficient Memory Management for Large Language Model Serving with PagedAttention
Woosuk Kwon, Zhuohan Li, Siyuan Zhuang +6
High throughput serving of large language models (LLMs) requires batching sufficiently many requests at a time. However, existing systems struggle because the key-value cache (KV c…
FlexGen: High-Throughput Generative Inference of Large Language Models with a Single GPU
Ying Sheng, Lianmin Zheng, Binhang Yuan +11
The high computational and memory requirements of large language model (LLM) inference make it feasible only with multiple high-end accelerators. Motivated by the emerging demand f…
On Optimizing the Communication of Model Parallelism
Yonghao Zhuang, Hexu Zhao, Lianmin Zheng +6
We study a novel and important communication pattern in large-scale model-parallel deep learning (DL), which we call cross-mesh resharding. This pattern emerges when the two paradi…
POET: Training Neural Networks on Tiny Devices with Integrated Rematerialization and Paging
Shishir G. Patil, Paras Jain, Prabal Dutta +2
Fine-tuning models on edge devices like mobile phones would enable privacy-preserving personalization over sensitive data. However, edge training has historically been limited to r…
Representing Long-Range Context for Graph Neural Networks with Global Attention
Zhanghao Wu, Paras Jain, Matthew A. Wright +3
Graph neural networks are powerful architectures for structured datasets. However, current methods struggle to represent long-range dependencies. Scaling the depth or width of GNNs…
Alpa: Automating Inter- and Intra-Operator Parallelism for Distributed Deep Learning
Lianmin Zheng, Zhuohan Li, Hao Zhang +9
Alpa automates model-parallel training of large deep learning (DL) models by generating execution plans that unify data, operator, and pipeline parallelism. Existing model-parallel…
PAC Best Arm Identification Under a Deadline
Brijen Thananjeyan, Kirthevasan Kandasamy, Ion Stoica +3
We study -PAC best arm identification, where a decision-maker must identify an -optimal arm with probability at least , while minimizing the number of arm pulls (…
Learning Competitive Equilibria in Exchange Economies with Bandit Feedback
Wenshuo Guo, Kirthevasan Kandasamy, Joseph E Gonzalez +2
The sharing of scarce resources among multiple rational agents is one of the classical problems in economics. In exchange economies, which are used to model such situations, agents…
ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training
Jianfei Chen, Lianmin Zheng, Zhewei Yao +4
The increasing size of neural network models has been critical for improvements in their accuracy, but device memory is not growing at the same rate. This creates fundamental chall…
Resource Allocation in Multi-armed Bandit Exploration: Overcoming Sublinear Scaling with Adaptive Parallelism
Brijen Thananjeyan, Kirthevasan Kandasamy, Ion Stoica +3
We study exploration in stochastic multi-armed bandits when we have access to a divisible resource that can be allocated in varying amounts to arm pulls. We focus in particular on…
Contrastive Code Representation Learning
Paras Jain, Ajay Jain, Tianjun Zhang +3
Recent work learns contextual representations of source code by reconstructing tokens from their context. For downstream semantic understanding tasks like summarizing code in Engli…
Checkmate: Breaking the Memory Wall with Optimal Tensor Rematerialization
Paras Jain, Ajay Jain, Aniruddha Nrusimha +5
We formalize the problem of trading-off DNN training time and memory requirements as the tensor rematerialization optimization problem, a generalization of prior checkpointing stra…
MLSys: The New Frontier of Machine Learning Systems
Alexander Ratner, Dan Alistarh, Gustavo Alonso +66
Machine learning (ML) techniques are enjoying rapidly increasing adoption. However, designing and implementing the systems that support ML models in real-world deployments remains…
Constrained Thompson Sampling for Wireless Link Optimization
Vidit Saxena, Joseph E. Gonzalez, Ion Stoica +2
Wireless communication systems operate in complex time-varying environments. Therefore, selecting the optimal configuration parameters in these systems is a challenging problem. Fo…