Few-shot learning continues to pose a challenge as it is inherently difficult for visual recognition models to generalize with limited labeled examples. When the training data is limited. the process of training and fine-tuning the model will be unstable and inefficient due to overfitting. In this paper. https://ashleyshomestores.shop/product-category/2-piece-desk/