docs(paper): 完善摘要中术语的完整表述和英文表达

修改中文摘要中"LLMs"为全称"Large Language Models"，并补充"SFT"的全称"Supervised Fine-Tuning"。英文摘要部分重写了开篇句式，使用更地道的学术表达，同时突出框架名称的强调格式。这些修改使论文摘要更加规范和专业。

paper/latex/chapters/abstract.tex

@@ -7,7 +7,7 @@
 
 % 中文摘要
 \begin{onecolabstract}
-	\noindent{}{\zihao{4}\textbf{摘要}}{\quad \songti \zihao{-4}大语言模型（LLMs）在通用代码生成任务中表现出色，但在处理包含专有知识的企业私有代码库时，其性能往往受限。针对此问题，本文提出并实现了一个基于文档驱动的自适应编码大模型微调框架。该框架的核心创新在于：首先，通过深度解析技术文档（Markdown格式），自动抽取信息并结合预设模板生成高质量的指令微调（SFT）训练语料；其次，利用参数高效微调技术（如Quantized Low-Rank Adaptation，QLoRA）对预训练的编码大模型（以qwen2.5为例）进行针对性优化，使其精准适配私有库特有的语法结构与编程范式；最后，整合了包括数据持久化（SQLite+TinyDB）、训练监控（TensorBoard）和交互式前端（Gradio）在内的完整工作流。实验结果表明，该框架能够有效提升大模型在私有库代码生成任务上的准确性和实用性，为实现企业级软件开发的智能化和高效化提供了一套自动化、可扩展的解决方案。
+	\noindent{}{\zihao{4}\textbf{摘要}}{\quad \songti \zihao{-4}大语言模型（Large Language Models，LLMs）在通用代码生成任务中表现出色，但在处理包含专有知识的企业私有代码库时，其性能往往受限。针对此问题，本文提出并实现了一个基于文档驱动的自适应编码大模型微调框架。该框架的核心创新在于：首先，通过深度解析技术文档（Markdown格式），自动抽取信息并结合预设模板生成高质量的指令微调（Supervised Fine-Tuning，SFT）训练语料；其次，利用参数高效微调技术（如量化低秩微调，Quantized Low-Rank Adaptation，QLoRA）对预训练的编码大模型（以qwen2.5为例）进行针对性优化，使其精准适配私有库特有的语法结构与编程范式；最后，整合了包括数据持久化（SQLite+TinyDB）、训练监控（TensorBoard）和交互式前端（Gradio）在内的完整工作流。实验结果表明，该框架能够有效提升大模型在私有库代码生成任务上的准确性和实用性，为实现企业级软件开发的智能化和高效化提供了一套自动化、可扩展的解决方案。
 	}\par
 	\noindent{}{\zihao{4}\textbf{关键词}}{\quad \zihao{-4}\songti 大语言模型；代码生成；模型微调；文档驱动
 	}\par
@@ -15,8 +15,8 @@
 
 % 英文摘要
 \begin{onecolabstract}
-	\noindent{}{ \zihao{4} \textbf{ABSTRACT}}{\quad \zihao{-4}Large Language Models (LLMs) excel in general code generation tasks, but their performance is often limited when handling enterprise private code repositories containing proprietary knowledge. To address this issue, this paper proposes and implements a document-driven adaptive fine-tuning framework for large code models. The core innovations of this framework are: first, by deeply parsing technical documentation (Markdown format), it automatically extracts information and combines it with preset templates to generate high-quality instruction fine-tuning (SFT) training data; second, it utilizes parameter-efficient fine-tuning techniques (such as Quantized Low-Rank Adaptation (QLoRA)) to specifically optimize a pre-trained large code model (taking qwen2.5 as an example), enabling it to accurately adapt to the specific syntax, structure, and programming paradigms of the private library; finally, it integrates a complete workflow including data persistence (SQLite+TinyDB), training monitoring (TensorBoard), and an interactive frontend (Gradio). Experimental results demonstrate that this framework can effectively improve the accuracy and practicality of large models in private library code generation tasks, and provide an automated, scalable solution for intelligent and efficient enterprise software development.
+	\noindent{}{ \zihao{4} \textbf{ABSTRACT}}{\quad \zihao{-4}While Large Language Models (LLMs) demonstrate impressive performance in general code generation, their efficacy often diminishes when dealing with enterprise private code repositories that contain proprietary knowledge. To address this challenge, this paper proposes and implements a **document-driven adaptive fine-tuning framework for large code models**. The framework introduces several key innovations: Firstly, it employs deep parsing techniques on technical documentation (in Markdown format) to automatically extract information and combine it with predefined templates, thereby generating high-quality Supervised Fine-Tuning (SFT) training data. Secondly, it leverages parameter-efficient fine-tuning techniques, such as Quantized Low-Rank Adaptation (QLoRA), to specifically optimize a pre-trained large code model (exemplified by Qwen2.5). This optimization enables the model to accurately adapt to the unique syntax, structures, and programming paradigms inherent in private libraries. Finally, the framework integrates a comprehensive workflow encompassing data persistence (using SQLite and TinyDB), training monitoring (via TensorBoard), and an interactive frontend (built with Gradio). Experimental results indicate that this framework significantly enhances the accuracy and practical utility of LLMs in private library code generation tasks, offering an automated and scalable solution for intelligent and efficient enterprise software development.
 	}\par
 	\noindent{}{ \zihao{4}\textbf{KEYWORDS}}{\quad \zihao{-4}Large Language Models; Code Generation; Model Fine-tuning; Document-Driven
 	}\par
-\end{onecolabstract}
+\end{onecolabstract}