[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-115459-en":3,"doc-seo-115459-105":31,"detail-sidebar-cat-0-en-105":92},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":21,"is_downloadable":21,"audit_status":21,"page_count":22,"language":23,"language_code":24,"site_id":25,"html_lang":24,"table_of_contents":26,"faqs":27,"seo_title":28,"seo_description":14,"update_tm":29,"read_time":30},115459,1099523885074,"Ivy","https://ap-avatar.wpscdn.com/davatar_9964176cb1d06d4a9deccf72a44ae3dc",6,"Technology","Generative AI with Python - 教学与应用路线图","Generative AI with Python系统梳理生成式人工智能的核心概念与主流技术路径，涵盖文本、图像、语音、翻译、图像描述、对话、结构化摘要与抽取、RAG语义检索、OCR与表格抽取、视频生成、工具调用、以及合成数据生成。内容进一步讨论多模态融合、推理模型、硬件与推理优化，并给出从软件开发到办公、设计营销、金融客服、工业与健康公共服务的落地场景，同时强调治理、可溯源性、评测与安全基准的重要性。","CONTENTS  \nGenerative AI with Python  \nGenerative AI  \nPython for AI  \nPreparing the Environment  \nText → Image  \nText → Speech  \nSpeech → Text  \nTranslation  \nImage Description  \nChat  \nSummary & Structured Extraction  \nRAG (Semantic Search)  \nOCR & Table Extraction  \nText/Image → Video  \nTool Calling  \nSynthetic Data Generation  \nGenerative AI and Python  \n[OceanofPDF.com](OceanofPDF.com)  \nGENERATIVE AI WITH PYTHON  \nEgon Kirchof Mazera  \nUndergraduate in Computer Systems (Universidade Federal Fluminense)  \nMaster in AI (Universidad de Valencia)  \nDisclaimer: The examples in this book are about Brazil my home country.  \n[OceanofPDF.com](OceanofPDF.com)  \nGENERATIVE AI  \nGenerative artificial intelligence is, at the same time, a technology and a new way of producing cultural and scientific material: models that learn patterns from data and then produce text, images, audio, video, code, and multimodal combinations with fluency increasingly close to that of humans.  \nAt the foundation, there are two families that have become ubiquitous:  \n-Attention-based transformers (for language and vision-language)  \n-Diffusion models (for image and video synthesis) Transformers dominate because they can “look at” long sequencesand align distant parts of the context, enabling translation, multistep reasoning, and integration with tools. Diffusion models shine in visual fidelity and style control.  \nToday, however, the real frontier is convergence: multimodal architectures that listen, see, and speak in real time, with latencies low enough to make continuous natural interfaces viable. The launch of GPT-4o as an “omni” model captured this leap: text, vision, and audio in a single system, with significantly better auditory and visual understanding and aggressive price/performance for practical use cases. This sensory integration is not cosmetic; it enables workflows where a user points the camera at an electrical panel, describes the problem by speaking, receives an explanation overlaid on the image, and asks for a simulation or step-by-step guidance.  \nThe state of the art is also being pushed by other actors, reinforcing the plurality of the ecosystem. Anthropic has raised the bar for reasoning and utility with Claude 3.5 Sonnet, surpassing previous generations of the company’s own models and introducing“computer use”, where the agent can operate interfaces and apps more autonomously. The combination of capability, cost, and speed is redefining what kinds of tasks are feasible to offload to AI in day-to-day work.  \nGoogle, in turn, has invested heavily in the Gemini family, especially Gemini 1.5 with very large context windows, capable of ingesting long documents, extensive codebases, or hours of video. This drastically changes the type of problem that can “fit inside the model’s head”, from large-scale legal analysis to auditing entire repositories.  \nOn the hardware side, NVIDIA has consolidated itself as a central pillar of the current wave, with data-center GPUs connected by NVLink and gigantic clusters that behave almost like a single logical GPU. Public demos have shown record throughput measured in tokens per second per user. Optimizations such as better kernels, quantization, speculative decoding, and reducedprecision formats like FP8 all combine to make serving large models cheaper and more scalable.  \nWe are also seeing the emergence of “reasoning models”, which emphasize overt reasoning steps, exploration of hypotheses, and integration of multimodal evidence rather than just fluent text. This is a sign of maturity in the direction of sustained usefulness rather than just impressive demos.  \nIn concrete use, the range of applications is already broad and profitable. In software, copilots write and refactor code, explain legacy systems, and generate tests. In office work, they help draft  \nemails, reports, and presentations. In design and marketing, they accelerate the creation of images, layouts, and variations of campaigns target","cbCaijP97EOzA02O","https://ap.wps.com/l/cbCaijP97EOzA02O","pdf",23079931,2,1,133,"English","en",105,"# Contents\n## Generative AI\n## Python for AI\n## Preparing the Environment\n## Text → Image\n## Text → Speech\n## Speech → Text\n## Translation\n## Image Description\n## Chat\n## Summary & Structured Extraction\n## RAG (Semantic Search)\n## OCR & Table Extraction\n## Text/Image → Video\n## Tool Calling\n## Synthetic Data Generation","[{\"question\":\"本书重点讲解哪些生成式AI能力？\",\"answer\":\"覆盖文本到图像、文本到语音、语音到文本、翻译、图像描述、对话、结构化摘要与抽取、语义检索（RAG）、OCR与表格抽取、文本/图像到视频、工具调用以及合成数据生成等能力。\"},{\"question\":\"为什么多模态融合被认为是生成式AI的“前沿”？\",\"answer\":\"文中指出，多模态架构能实时“听、看、说”，并通过低延迟实现连续自然交互，使相机识别、语音描述问题、图像叠加解释及进一步仿真或分步指导等工作流成为可能。\"},{\"question\":\"本书如何讨论生成式AI在落地中的治理与安全？\",\"answer\":\"强调风险分级治理、文档与透明度、解释性与鲁棒性，以及可溯源性（如C2PA）与评测工具链（基准测试、红队、领域测试集）来监控幻觉、偏差与安全漏洞，并评估与外部工具组合后的行为。\"}]","Generative AI with Python - 教学与应用路线图 | PDF",1785473646,335,{"code":4,"msg":32,"data":33},"ok",{"site_id":25,"language":24,"slug":34,"title":13,"keywords":35,"description":14,"schema_data":36,"social_meta":87,"head_meta":89,"extra_data":91,"updated_unix":29},"generative-ai-with-python-teaching-and-application-roadmap","",{"@graph":37,"@context":86},[38,54,69],{"@type":39,"itemListElement":40},"BreadcrumbList",[41,45,48,51],{"item":42,"name":43,"@type":44,"position":21},"https://docshare.wps.com","Home","ListItem",{"item":46,"name":47,"@type":44,"position":20},"https://docshare.wps.com/document/","Document",{"item":49,"name":12,"@type":44,"position":50},"https://docshare.wps.com/document/technology/",3,{"item":52,"name":13,"@type":44,"position":53},"https://docshare.wps.com/document/generative-ai-with-python-teaching-and-application-roadmap/115459/",4,{"url":52,"name":13,"@type":55,"author":56,"headline":13,"publisher":58,"fileFormat":61,"inLanguage":24,"description":14,"dateModified":62,"datePublished":63,"encodingFormat":61,"isAccessibleForFree":64,"interactionStatistic":65},"DigitalDocument",{"name":9,"@type":57},"Person",{"url":42,"name":59,"@type":60},"DocShare","Organization","application/pdf","2026-08-16","2026-07-31",true,{"@type":66,"interactionType":67,"userInteractionCount":20},"InteractionCounter",{"@type":68},"ViewAction",{"@type":70,"mainEntity":71},"FAQPage",[72,78,82],{"name":73,"@type":74,"acceptedAnswer":75},"本书重点讲解哪些生成式AI能力？","Question",{"text":76,"@type":77},"覆盖文本到图像、文本到语音、语音到文本、翻译、图像描述、对话、结构化摘要与抽取、语义检索（RAG）、OCR与表格抽取、文本/图像到视频、工具调用以及合成数据生成等能力。","Answer",{"name":79,"@type":74,"acceptedAnswer":80},"为什么多模态融合被认为是生成式AI的“前沿”？",{"text":81,"@type":77},"文中指出，多模态架构能实时“听、看、说”，并通过低延迟实现连续自然交互，使相机识别、语音描述问题、图像叠加解释及进一步仿真或分步指导等工作流成为可能。",{"name":83,"@type":74,"acceptedAnswer":84},"本书如何讨论生成式AI在落地中的治理与安全？",{"text":85,"@type":77},"强调风险分级治理、文档与透明度、解释性与鲁棒性，以及可溯源性（如C2PA）与评测工具链（基准测试、红队、领域测试集）来监控幻觉、偏差与安全漏洞，并评估与外部工具组合后的行为。","https://schema.org",{"og:url":52,"og:type":88,"og:title":13,"og:site_name":59,"og:description":14},"article",{"robots":90,"canonical":52},"index,follow",{"doc_id":7,"site_id":25},{"code":4,"msg":5,"data":93},[94,98,102,106,111,114,119,124,129,132,136],{"id":21,"doc_module":4,"doc_module_name":47,"category_name":95,"show_sort_weight":96,"slug":97},"Story & Novel",90,"story-novel",{"id":20,"doc_module":4,"doc_module_name":47,"category_name":99,"show_sort_weight":100,"slug":101},"Literature",80,"literature",{"id":53,"doc_module":4,"doc_module_name":47,"category_name":103,"show_sort_weight":104,"slug":105},"Exam",70,"exam",{"id":107,"doc_module":4,"doc_module_name":47,"category_name":108,"show_sort_weight":109,"slug":110},5,"Comic",60,"comic",{"id":11,"doc_module":4,"doc_module_name":47,"category_name":12,"show_sort_weight":112,"slug":113},50,"technology",{"id":115,"doc_module":4,"doc_module_name":47,"category_name":116,"show_sort_weight":117,"slug":118},7,"Healthcare",40,"healthcare",{"id":120,"doc_module":4,"doc_module_name":47,"category_name":121,"show_sort_weight":122,"slug":123},8,"Research & Report",30,"research-report",{"id":125,"doc_module":4,"doc_module_name":47,"category_name":126,"show_sort_weight":127,"slug":128},9,"Religion & Spirituality",20,"religion-spirituality",{"id":127,"doc_module":4,"doc_module_name":47,"category_name":130,"show_sort_weight":127,"slug":131},"World Cup","world-cup",{"id":133,"doc_module":4,"doc_module_name":47,"category_name":134,"show_sort_weight":133,"slug":135},10,"Lifestyle","lifestyle",{"id":137,"doc_module":4,"doc_module_name":47,"category_name":138,"show_sort_weight":107,"slug":139},19,"General","general"]