基于Lora模型的傳統(tǒng)服飾數(shù)字化傳承與創(chuàng)新設(shè)計
——以蒙古族服飾為例
打開文本圖片集
Abstract：Thefocusofthisstudyisonthetextualresearchanddigitizationoftraditionalclothing，trainingtoformastableLora model.Buildingonthis，itcombinesspecificuseranddesignerneedstocreatedesigns withconsistentstyles.Thestudydigitizes typical elementsof Mongoliantraditional Costumeandanalyzes thecolorsof typicalMongolian garments，processing them digitally toasistindivergentdesign.Itestablises workflowstailoredtovarious design stylesandtests theimagegeneration efectsofthemodelacross diferentpaths.Theresultsindicatethatusingacomprehensivedatasetoftraditionalclothingand itscolors，astableLoramodelcanbeformedthroughdeplearning.Thedifusionmodel，basedonthisgeneratesimagesand designs i multiplestyles，efectivelyachieving cultural heritageand innovativedesign migrationoftraditionalclothing.This providesareferenceforfashiondesignersandresearchers inthefashionindustrytouseAIGCtechnologyfortheinheritaceand innovation of traditional Chinese clothing.
KeyWords：LoraModel;MachineLearning;CostumeColor;MongolianAttire
近年來，隨著AI技術(shù)的發(fā)展，越來越多的研究者探索并擴展著數(shù)字藝術(shù)與設(shè)計創(chuàng)意需求的邊界，擴散模型、去噪擴散概率模型、潛空間擴散模型等不斷地迭代，不斷地提高擴散模型的圖像生成效率。（剩余6485字）