Research on the Generation Method of Liblib Non-Heritage Qiang Embroidery Sample Based on LoRA Model Fine-Tuning

This research explores the generation of Qiang embroidery patterns using LoRA model fine-tuning. By analyzing the unique style factors of Qiang embroidery, multiple style-specific LoRA models were constructed to precisely control pattern generation through text prompts. The results show that the generated patterns meet the aesthetic standards of Qiang embroidery, with high consistency with the intended designs. Subjective evaluations by experts indicate that the patterns have strong cultural uniqueness and visual expressiveness. This study provides a new technical approach for the digital inheritance and innovation of Qiang embroidery, reducing the workload of artisans and expanding the application space for intangible cultural heritage.