LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues

• URL: http://arxiv.org/abs/2507.13681v1
• Date: Fri, 18 Jul 2025 06:12:08 GMT
• Title: LoopServe: An Adaptive Dual-phase LLM Inference Acceleration System for Multi-Turn Dialogues
• Authors: Haoyang Li, Zhanchao Xu, Yiming Li, Xuejia Chen, Darian Li, Anxin Tian, Qingfa Xiao, Cheng Deng, Jun Wang, Qing Li, Lei Chen, Mingxuan Yuan,
• Abstract summary: LoopServe is an adaptive dual-phase inference framework for large language models in multi-turn dialogues.<n>It consistently achieves superior effectiveness compared to existing baselines.<n>It significantly accelerates inference across a wide range of long-context dialogue tasks.
• Score: 55.332261548836485
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Multi-turn dialogues are essential in many real-world applications of large language models, such as chatbots and virtual assistants. As conversation histories become longer, existing large language models face increasing computational and memory challenges, which hinder their ability to provide efficient and responsive interactions. Most current acceleration methods either compress the context or optimize key value caching, but they often rely on fixed or position-based heuristics that do not adapt well to the dynamic and unpredictable patterns found in actual multi-turn conversations. In this paper, we present LoopServe, an adaptive dual-phase inference acceleration framework for large language models in multi-turn dialogues. LoopServe introduces two main innovations. First, it performs online sparsification during the prefilling phase by dynamically selecting the most important parts of the attention matrix for each new input. Second, it uses progressive key value compression during decoding by adaptively maintaining a relevant and efficient cache based on the most recently generated output tokens. We also propose a \href{https://huggingface.co/datasets/TreeAILab/Multi-turn_Long-context_Benchmark_for_LLMs}{new benchmark} with eleven multi-turn datasets that reflect realistic query positions and conversational dependencies. Extensive experiments demonstrate that LoopServe consistently achieves superior effectiveness compared to existing baselines and significantly accelerates LLM inference across a wide range of long-context dialogue tasks.

Related papers

• ContextCache: Context-Aware Semantic Cache for Multi-Turn Queries in Large Language Models [33.729482204460815]
  This demonstration introduces ContextCache, a context-aware semantic caching system for multi-turn dialogues.<n> ContextCache employs a two-stage retrieval architecture that first executes vector-based retrieval on the current query to identify potential matches and then integrates current and historical dialogue representations through self-attention mechanisms for precise contextual matching.<n> cached responses exhibit approximately 10 times lower latency than direct LLM invocation, enabling significant computational cost reductions for conversational applications.
  arXiv  Detail & Related papers  (2025-06-28T07:25:12Z)
• MEM1: Learning to Synergize Memory and Reasoning for Efficient Long-Horizon Agents [84.62985963113245]
  We introduce MEM1, an end-to-end reinforcement learning framework that enables agents to operate with constant memory across long multi-turn tasks.<n>At each turn, MEM1 updates a compact shared internal state that jointly supports memory consolidation and reasoning.<n>We show that MEM1-7B improves performance by 3.5x while reducing memory usage by 3.7x compared to Qwen2.5-14B-Instruct on a 16-objective multi-hop QA task.
  arXiv  Detail & Related papers  (2025-06-18T19:44:46Z)
• Streaming Video Understanding and Multi-round Interaction with Memory-enhanced Knowledge [57.01131456894516]
  Current video understanding models struggle with processing long video sequences, supporting multi-turn dialogues, and adapting to real-world dynamic scenarios.<n>We propose StreamChat, a training-free framework for streaming video reasoning and conversational interaction.<n>Our framework incorporates a parallel system scheduling strategy that enhances processing speed and reduces latency, ensuring robust performance in real-world applications.
  arXiv  Detail & Related papers  (2025-01-23T08:33:10Z)
• Data-Centric Improvements for Enhancing Multi-Modal Understanding in Spoken Conversation Modeling [13.628984890958314]
  We introduce a data-centric customization approach for efficiently enhancing multimodal understanding in conversational speech modeling.<n>Our approach achieves state-of-the-art performance on the Spoken-SQuAD benchmark, using only 10% of the training data with open-weight models.<n>We also introduce ASK-QA, the first dataset for multi-turn spoken dialogue with ambiguous user requests and dynamic evaluation inputs.
  arXiv  Detail & Related papers  (2024-12-20T15:43:09Z)
• InternLM-XComposer2.5-OmniLive: A Comprehensive Multimodal System for Long-term Streaming Video and Audio Interactions [104.90258030688256]
  This project introduces disentangled streaming perception, reasoning, and memory mechanisms, enabling real-time interaction with streaming video and audio input.<n>This project simulates human-like cognition, enabling multimodal large language models to provide continuous and adaptive service over time.
  arXiv  Detail & Related papers  (2024-12-12T18:58:30Z)
• QPO: Query-dependent Prompt Optimization via Multi-Loop Offline Reinforcement Learning [58.767866109043055]
  We introduce Query-dependent Prompt Optimization (QPO), which iteratively fine-tune a small pretrained language model to generate optimal prompts tailored to the input queries.<n>We derive insights from offline prompting demonstration data, which already exists in large quantities as a by-product of benchmarking diverse prompts on open-sourced tasks.<n> Experiments on various LLM scales and diverse NLP and math tasks demonstrate the efficacy and cost-efficiency of our method in both zero-shot and few-shot scenarios.
  arXiv  Detail & Related papers  (2024-08-20T03:06:48Z)
• LoongServe: Efficiently Serving Long-Context Large Language Models with Elastic Sequence Parallelism [12.521026493432181]
  Existing large language models (LLMs) cannot efficiently serve variable-length requests in different phases. We propose a new parallelism paradigm, elastic sequence parallelism (ESP), to adapt to the variance between different requests and phases. LoongServe improves the maximum throughput by up to 3.85$times$ compared to the chunked prefill and 5.81$times$ compared to the prefill-decoding disaggregation.
  arXiv  Detail & Related papers  (2024-04-15T07:45:04Z)
• DialCLIP: Empowering CLIP as Multi-Modal Dialog Retriever [83.33209603041013]
  We propose a parameter-efficient prompt-tuning method named DialCLIP for multi-modal dialog retrieval. Our approach introduces a multi-modal context generator to learn context features which are distilled into prompts within the pre-trained vision-language model CLIP. To facilitate various types of retrieval, we also design multiple experts to learn mappings from CLIP outputs to multi-modal representation space.
  arXiv  Detail & Related papers  (2024-01-02T07:40:12Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.