[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-651109-105":59,"doc-detail-651109-en":124},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":117,"head_meta":119,"extra_data":121,"updated_unix":123},105,"en","fine-tuning-stable-diffusion-for-generating-2d-floor-plans-using-prompt-templates","Fine-Tuning Stable Diffusion for Generating 2D Floor Plans - Using Prompt Templates","","Automated generation of 2D floor plans is essential for architectural design, but models must balance precision with responsiveness to user-defined constraints. Stable Diffusion can produce high-quality images yet lacks inherent layout understanding, while graph neural networks capture object relationships but are not generative or text-driven. This study fine-tunes Stable Diffusion 1.5 using a custom floor-plan dataset and structured prompt templates to limit free-form creativity. Results indicate that prompt engineering and dataset preprocessing materially improve output quality and robustness, supporting diffusion-based architecture for structured layout creation.",{"@graph":69,"@context":116},[70,84,99],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/fine-tuning-stable-diffusion-for-generating-2d-floor-plans-using-prompt-templates/651109/",{"url":83,"name":65,"@type":85,"author":86,"headline":65,"publisher":89,"fileFormat":92,"inLanguage":63,"description":67,"dateModified":93,"datePublished":93,"encodingFormat":92,"isAccessibleForFree":94,"interactionStatistic":95},"DigitalDocument",{"name":87,"@type":88},"bocahgabuttt","Person",{"url":74,"name":90,"@type":91},"DocShare","Organization","application/pdf","2026-10-24",true,{"@type":96,"interactionType":97,"userInteractionCount":4},"InteractionCounter",{"@type":98},"ViewAction",{"@type":100,"mainEntity":101},"FAQPage",[102,108,112],{"name":103,"@type":104,"acceptedAnswer":105},"Why are diffusion models like Stable Diffusion useful for floor plan generation, and what is their limitation?","Question",{"text":106,"@type":107},"Stable Diffusion generates high-quality images, but it does not inherently understand structured layouts like floor plans. This motivates combining it with approaches that can represent spatial structure more explicitly.","Answer",{"name":109,"@type":104,"acceptedAnswer":110},"How do graph neural networks (GNNs) contribute, and why are they not sufficient alone?",{"text":111,"@type":107},"GNNs encode relational data by representing floor plan objects as nodes and connections as edges. However, they are not generative and cannot directly process textual inputs to produce floor plans.",{"name":113,"@type":104,"acceptedAnswer":114},"What method does the paper propose to improve output quality for 2D floor plans?",{"text":115,"@type":107},"The paper fine-tunes Stable Diffusion 1.5 on a custom floor-plan dataset using structured prompt templates. It also emphasizes prompt engineering and dataset preprocessing to constrain creativity and improve accuracy and error tolerance.","https://schema.org",{"og:url":83,"og:type":118,"og:title":65,"og:site_name":90,"og:description":67},"article",{"robots":120,"canonical":83},"index,follow",{"doc_id":122,"site_id":62},651109,1792862884,{"code":4,"msg":5,"data":125},{"doc_id":122,"user_id":126,"nickname":87,"user_avatar":127,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":128,"file_id":129,"file_url":130,"file_type":131,"file_size":132,"view_count":4,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":46,"language":133,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":134,"faqs":135,"seo_title":136,"seo_description":67,"update_tm":123,"read_time":31},962094248839,"https://ap-avatar.wpscdn.com/davatar_994ba38a5ba835b3df7d355c54d3ed8d","Fine-Tuning Stable Diffusion for Generating 2D Floor Plans  \nUsing Prompt Templates  \nAhmed Mostafa, Omar Amir , Ali M. Mohamed  and Marwa O. Al Enany∗   \nHigher Institute of Computer Science and Information Systems,  \nOctober 6 University, Culture and Science City, Egypt  \n∗ Corresponding author: Marwa O. Al Enany ([marwaanny33@gmail.com](marwaanny33@gmail.com))  \nSubmitted: 02 Feb 2025 Accepted: 20 May 2025 Published: 30 Sep 2025 License: CC BY-NC 4 .0 cbn  \nAbstract Automated generation of 2D floor plans is crucial for architectural design, requiring models to balance precision and adaptability to user-defined specifications. Diffusion models, like Stable Diffusion, excel at generating high-quality images but lack an intrinsic understanding of structured layouts such as floor plans. Conversely, Graph Neural Networks (GNNs) are adept at encoding relational data, representing floor plan objects as nodes and their connections as edges, but they are not generative or capable of processing textual inputs. In this work, we fine-tune Stable Diffusion 1.5 on a custom dataset of floor plans, leveraging structured prompt templates to constrain the model’s creativity and guide it toward generating concise, error-tolerant outputs. This research suggests integrating the generative capabilities of diffusion models with the representational strengths of GNNs to overcome inherent challenges in diffusion models, like their inability to explicitly encode spatial relationships. This integration could expand the capabilities of these models, empowering them to comprehend and produce structured layouts more effectively. While computational constraints limited our exploration of this hybrid architecture, our results demonstrate that prompt engineering and dataset preprocessing significantly improve the output quality. This study highlights the potential for generative models in architectural tasks and lays the groundwork for integrating logical reasoning into diffusion-based architectures.  \nKeywords: graph neural networks (GNNs), diffusion model, latent diffusion, floorplan representation.  \n1. Introduction  \nHuman culture deeply intertwines with the history of architecture and architectural drawings, reflecting the ability to conceptualize and design living spaces. Architectural drawings, meticulously crafted by hand and archived on paper or as scanned raster images, continued until the second half of the 20th century. Even with the widespread adoption of computer-aided design (CAD) software [7], most architectural drawings remain primarily distributed in image formats, limiting the depth of information that can be extracted and analyzed.  \nThe challenge of generating floor plans that meet specific user requirements has long been a complex task for architects and designers [33] . Traditional methods rely heavily on manual design processes, requiring extensive expertise and time-consuming iterations. This process is inherently complex and requires meticulous attention to detail. Architects and engineers must consider a myriad of factors, including spatial relationships, structural integrity, building codes, and client preferences. To ensure both functionality  \nMachine GRAPHICS & VISION 34(3):77–95, 2025 . DOI: 10.22630/MGV.2025.34.3.4 .  \nand compliance with regulatory standards, architects and engineers must precisely configure each element, from the placement of walls, doors, and windows to the allocation of rooms and incorporation of utilities. Moreover, the manual approach to floor plan creation poses challenges in efficiency and productivity. Engineers must painstakingly adjust dimensions, realign components, and ensure consistency across different sections of the plan. This meticulous work not only prolongs project timelines but also redirects valuable resources towards less innovative design aspects. The pressure to deliver accurate and high-quality floor plans under tight deadlines further exacerbates the strain on professionals in the","cbCaigRLvgWKg29z","https://ap.wps.com/l/cbCaigRLvgWKg29z","pdf",999033,"English","# Introduction\n## Background and motivation\n## Challenges of manual floor plan generation\n## Need for automated solutions\n## Related AI and ML approaches","[{\"question\":\"Why are diffusion models like Stable Diffusion useful for floor plan generation, and what is their limitation?\",\"answer\":\"Stable Diffusion generates high-quality images, but it does not inherently understand structured layouts like floor plans. This motivates combining it with approaches that can represent spatial structure more explicitly.\"},{\"question\":\"How do graph neural networks (GNNs) contribute, and why are they not sufficient alone?\",\"answer\":\"GNNs encode relational data by representing floor plan objects as nodes and connections as edges. However, they are not generative and cannot directly process textual inputs to produce floor plans.\"},{\"question\":\"What method does the paper propose to improve output quality for 2D floor plans?\",\"answer\":\"The paper fine-tunes Stable Diffusion 1.5 on a custom floor-plan dataset using structured prompt templates. It also emphasizes prompt engineering and dataset preprocessing to constrain creativity and improve accuracy and error tolerance.\"}]","Fine-Tuning Stable Diffusion for Generating 2D Floor Plans - Using Prompt Templates | PDF"]