从大型到超小型：面向高性价比 LLM 的端到端优化
From Large to Super-Tiny: End-to-End Optimization for Cost-Efficient LLMs
AI总结：
本文提出三阶段端到端 LLM 部署流水线，通过函数调用原型、知识迁移及量化剪枝，将模型压缩至 0.4B，在降低成本与延迟的同时保持性能。
摘要：
大型语言模型（LLM）通过优化传统自然语言处理（NLP）工作流并推动其集成到各类系统中，显著推动了人工智能的发展。许多此类 NLP 系统（包括我们的系统）都直接采用 LLM。然而，这种方式要么导致高昂成本，要么在微调后产生次优性能。本文提出一种包含原型设计、知识迁移和模型压缩的三阶段高性价比端到端 LLM 部署流水线，以有效解决基于 LLM 的框架中的成本—性能困境。其高成本效益不仅体现在简化系统复杂性，并在结果中生成性能增强、成本降低的超小型在线模型，还体现在应对项目开发过程中的开发周期约束、缺乏大量高质量数据以及计算资源有限等问题。第一阶段，我们通过将复杂任务转换为基于函数调用的 LLM 驱动流水线，构建性能最优的原型系统，并将其作为教师模型生成高质量数据。第二阶段，我们结合拒绝采样微调、强化学习和知识蒸馏等技术，将知识迁移到 0.5B 学生模型，以最低成本提供有效性能。最后阶段，我们通过量化和剪枝将模型进一步压缩至 0.4B，实现超低延迟和成本。大量实验结果以及该框架的模块化设计表明，其具备跨领域能力，并有可能应用于其他 NLP 领域。
英文摘要：
Large Language Models (LLMs) have significantly advanced artificial intelligence by optimizing traditional Natural Language Processing (NLP) workflows, facilitating their integration into various systems. Many such NLP systems, including ours, directly incorporate LLMs. However, this approach either results in expensive costs or yields suboptimal performance after fine-tuning. In this paper, we introduce a three-stage cost-efficient end-to-end LLM deployment pipeline, comprising prototyping, knowledge transfer, and model compression, to effectively tackle the cost-performance dilemma in LLM-based frameworks. Its high cost-efficiency is manifested not only in simplifying system complexity and producing super-tiny online models with enhanced performance and reduced costs in the results, but also in addressing development cycle constraints, the lack of extensive high-quality data, and limited computational resources during the project development process. In the first stage, we construct an optimal performance prototype system by transforming complex tasks into a function call-based LLM-driven pipeline, which serves as a teacher model to generate high-quality data. In the second stage, we combine techniques like rejection sampling fine-tuning, reinforcement learning, and knowledge distillation to transfer knowledge to 0.5B student models, delivering effective performance at minimal cost. In the final stage, we further compress models to 0.4B via quantization and pruning, achieving ultra-low latency and cost. Extensive experimental results and the framework's modular design suggest cross-domain capabilities and potential applicability in other NLP areas.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
