-
Notifications
You must be signed in to change notification settings - Fork 3
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
6 changed files
with
29 additions
and
45 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
16 changes: 8 additions & 8 deletions
16
contents/20250214-summarize-article-by-llm-inference/cn.summary.json
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -1,29 +1,29 @@ | ||
{ | ||
"articleId": "summary-article-by-llm-inference", | ||
"articleId": "summarize-article-by-llm-inference", | ||
"lang": "cn", | ||
"hash": "85c6edcc02b719c8ad5713c671bb50898b02ed4abfc5397cabeedbea1a318137", | ||
"hash": "7c12244a86d7ff5aed37ec8fb7bdcc88d8ab56dca1300b75e109bab2944d146d", | ||
"summaries": [ | ||
{ | ||
"summaries": [ | ||
"作者批评Azure AI语言服务的中文摘要效果差,转而通过Azure AI平台部署DeepSeek R1模型,设置特定Prompt后取得优秀效果。对比原服务机械化的关键词堆砌,R1能生成流畅自然的百字总结,并展示多篇文章改进案例。尝试本地部署DeepSeek R1 8B模型时发现小模型存在格式控制和语义理解缺陷,需通过Prompt工程调整。实践揭示LLM应用与传统软件开发的本质差异:业务逻辑内嵌于模型而非代码,提示词成为核心接口,这种黑箱特性带来结果不可预测性,但AI确实解决了传统方法难以实现的语义处理问题。" | ||
"作者批评Azure AI Language Service的中文摘要效果差,转而使用Azure部署的DeepSeek R1模型获得显著提升。通过设计精准提示词实现高质量文本摘要,成功应用于个人博客所有文章。尝试本地运行DeepSeek R1 8B版本时,发现小模型存在输出格式不稳定等问题。文章对比了传统编程与AI开发的差异,指出大模型将业务逻辑隐入黑箱带来的不确定性,但认可其突破传统技术限制的价值,期待未来探索更多适用场景。" | ||
], | ||
"metadata": { | ||
"summarizer": "azure-ai", | ||
"model": "DeepSeek-R1" | ||
}, | ||
"endTime": "2025-02-15T02:08:28.839Z", | ||
"startTime": "2025-02-15T02:07:51.243Z" | ||
"endTime": "2025-02-15T02:51:03.982Z", | ||
"startTime": "2025-02-15T02:49:57.481Z" | ||
}, | ||
{ | ||
"summaries": [ | ||
"<think>\n\n# 如何将AI应用于实际场景:我的初步体验\n\n作为一个AI开发者,我最近开始探索如何将大型语言模型(LLM)应用于实际场景。这次经历让我对LLM的使用流程有了更深刻的理解,同时也意识到其与传统软件工程的显著区别。\n\n在传统软件开发中,业务逻辑是通过代码精确地实现的。无论需求多么复杂,最终都能在代码中找到对应的逻辑。而使用LLM时,整个过程更像是在编写一段提示词,供模型根据上下文生成所需输出。这种模式带来了巨大的灵活性,但同时也增加了不确定性。\n\n我选择了一个特定的模型,并部署它到我的应用中。设计Prompt的过程是关键,因为它直接影响最终的输出结果。不同的Prompt可能会产生完全不同的回复,即使使用相同的输入,也可能得到相似的答案。这让我意识到,编写有效的Prompt是一门艺术,而非科学。\n\n在实际应用中,我遇到了几个挑战。首先,如何确保模型能够准确理解并处理特定的业务规则?其次,如何处理潜在的错误或不确定性?最后,如何将LLM集成到现有的系统中,并与其他组件协同工作。\n\n通过这次体验,我认识到以下几个关键点:\n\n1. **模型选择**:不同模型有不同的特点和限制,需要根据具体需求进行权衡。例如,对于需要高精度的应用,选择训练数据量多的大型模型会更合适。\n\n2. **Prompt设计**:设计有效的Prompt是非常挑战性的。这不仅需要对业务知识的深刻理解,还需要掌握如何引导模型生成所需输出。\n\n3. **错误处理和可靠性**:LLM虽然灵活,但在处理复杂问题时可能会出错。因此,必须有机制来检测和纠正这些错误,或至少在不影响关键业务流程的情况下进行容错处理。\n\n4. **集成与维护**:将LLM集成到现有系统中需要考虑硬件资源、API接口的设计、以及如何持续优化模型性能。这对技术团队提出了更高的要求,尤其是在代码维护和更新方面。\n\n尽管面临诸多挑战,我认为LLM在解决复杂问题和提供创新的可能性方面具有无限潜力。通过不断的探索和实践,我希望能够将这些技术应用于更多实际场景,推动业务的发展与创新。\n\n# 结语\n\n此次使用大型语言模型的经历让我深刻体会到其独特之处。虽然与传统软件开发不同,但LLM为解决复杂问题提供了全新的思路和方法。我期待未来能够将这些技术应用得更加广泛,推动更多领域的进步。" | ||
"这篇文章探讨了作者在“最好的年龄”如何度过这一阶段,涉及身体变化、工作环境不确定性以及内心的挣扎。总结起来,作者通过对工作和游戏的多角度描写,揭示了成熟与不稳定之间的矛盾,并鼓励读者反思自己的生活态度和价值观选择。文章语言流畅,情感真挚,但大模型在生成内容时未能完全满足格式、字数和语言要求,这表明在特定场景下,大模型的使用仍需优化。同时,作者也提到了对传统软件与LLM开发差异的思考,以及硬件升级建议,以支持更大的模型运行。" | ||
], | ||
"metadata": { | ||
"summarizer": "ollama", | ||
"model": "deepseek-r1:8b" | ||
}, | ||
"endTime": "2025-02-15T02:08:48.295Z", | ||
"startTime": "2025-02-15T02:08:28.841Z" | ||
"endTime": "2025-02-15T02:51:34.865Z", | ||
"startTime": "2025-02-15T02:51:03.983Z" | ||
} | ||
] | ||
} |
File renamed without changes
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters