Publications (9)
TurboSpec: Closed-loop Speculation Control System for Optimizing LLM Serving Goodput
Xiaoxuan Liu, Jongseok Park, Langxiang Hu +10
Large Language Model (LLM) serving systems batch concurrent user requests to achieve efficient serving. However, in real-world deployments, such inter-request parallelism from batc…
Learned Token Pruning for Transformers
Sehoon Kim, Sheng Shen, David Thorsley +4
Deploying transformer models in practice is challenging due to their inference cost, which scales quadratically with input sequence length. To address this, we present a novel Lear…
Jenga: Effective Memory Management for Serving LLM with Heterogeneity
Chen Zhang, Kuntai Du, Shu Liu +10
Large language models (LLMs) are widely used but expensive to run, especially as inference workloads grow. To lower costs, maximizing the request batch size by managing GPU memory…
APEX: An Extensible and Dynamism-Aware Simulator for Automated Parallel Execution in LLM Serving
Yi-Chien Lin, Woosuk Kwon, Ronald Pineda +1
Efficiently serving Large Language Models (LLMs) requires selecting an optimal parallel execution plan, balancing computation, memory, and communication overhead. However, determin…
Gemma 3 Technical Report
Gemma Team, Aishwarya Kamath, Johan Ferret +209
We introduce Gemma 3, a multimodal addition to the Gemma family of lightweight open models, ranging in scale from 1 to 27 billion parameters. This version introduces vision underst…
Gemma 2: Improving Open Language Models at a Practical Size
Gemma Team, Morgane Riviere, Shreya Pathak +195
In this work, we introduce Gemma 2, a new addition to the Gemma family of lightweight, state-of-the-art open models, ranging in scale from 2 billion to 27 billion parameters. In th…
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…
Nimble: Lightweight and Parallel GPU Task Scheduling for Deep Learning
Woosuk Kwon, Gyeong-In Yu, Eunji Jeong +1
Deep learning (DL) frameworks take advantage of GPUs to improve the speed of DL inference and training. Ideally, DL frameworks should be able to fully utilize the computation power…
A Fast Post-Training Pruning Framework for Transformers
Woosuk Kwon, Sehoon Kim, Michael W. Mahoney +3
Pruning is an effective way to reduce the huge inference cost of Transformer models. However, prior work on pruning Transformers requires retraining the models. This can add high t…