当前位置:首页 > 报告详情

让法学硕士 (LLM) 写题目:复合 AI 管道中的 DSPy 简介.pdf

上传人: Fl****zo 编号:719040 2025-06-22 41页 4.56MB

1、Forward-looking StatementThis presentation has been prepared for informational purposes only.The information set forth herein does not purport to be complete or contain all relevant information.Statements contained herein are made as of the date of this presentation unless stated otherwise.This pres

2、entation and the accompanying oral commentary may contain forward-looking statements.In some cases,forward-looking statements can be identified by terms such as“may”,“will”,“should”,“expects”,“plans”,“anticipates”,“could”,“intends”,“projects”,“believes”,“estimates”,“predicts”,or“continue”,or the neg

3、ative of these words or other similar terms or expressions that concern Databricks expectations,strategy,plans,or intentions.Forward-looking statements are based on information available at the time those statements are made and are inherently subject to risks and uncertainties that could cause actu

4、al results to differ materially from those expressed in or suggested by the forward-looking statements.Forward-looking statements should not be read as a guarantee of future performance or outcomes.Except as required by law,Databricks does not undertake any obligation to publicly update or revise an

5、y forward-looking statement,whether as a result of new information,future developments or otherwise.2Complete Your SurveysYou will receive a survey for each session attendedOpen the Databricks Events app and select“My Surveys”from the menuSurveys can also be submitted in the Attendee Portal3Your fee

6、dback has a direct impact on Data+AI Summit contentLet the LLM Write the PromptsAn Intro to DSPy in Compound AI PipelinesDrew BreunigTBD“Some people,when confronted with a problem,think I know,Ill use regular expressions.Now they have two problems.”5Jamie Zawinsky,somewhat apocryphally“Some people,w

word格式文档无特别注明外均可编辑修改,预览文件经过压缩,下载原文更清晰!
三个皮匠报告文库所有资源均是客户上传分享,仅供网友学习交流,未经上传用户书面授权,请勿作商用。
本文主要介绍了DSPy在复合AI管道中的应用,强调了其在优化提示语(prompts)方面的优势。 - **关键数据**:使用DSPy后,数据准确率从60.7%提升至82.0%,并使用了14行代码管理约700个token的提示语。 - **关键点**: 1. **提示语的优势与局限**:易于描述任务,但存在性能差异和组合边缘情况的问题。 2. **DSPy介绍**:将任务从LLM提示语中解耦,定义明确的输入输出,便于优化和模型切换。 3. **优化与模型切换**:DSPy支持多种算法和模型,易于优化和迁移,如从Qwen 3 0.6B到Phi-4-Mini 3.8B。 4. **实践案例**:通过DSPy优化地理位置信息合并(conflation)任务,提高了准确率。 5. **未来方向**:探索新优化器、微调、多阶段模块和工具整合。 - **结论**:DSPy使任务定义清晰,易于在复杂管道中管理,并支持持续优化,有助于保持竞争力。
"提升效率,DSPy如何帮你?" "地图数据优化,你试过Overture吗?" "告别复杂提示,体验编程新方式!"
客服
商务合作
小程序
服务号
折叠