SlimageNet64

SlimageNet64是一个为持续少量学习（CFSL）设计的紧凑数据集，由爱丁堡大学信息学院创建。该数据集包含1000个类别的200张图片，总计200000张64x64像素的RGB图像。SlimageNet64通过保留原始ImageNet的所有类别但减少每个类别的样本数量，旨在为CFSL提供一个高效且具有挑战性的基准。该数据集适用于研究样本效率、灾难性遗忘、神经网络的连续适应以及元学习系统的持续学习能力，特别适合在资源受限的环境中进行实验。

SlimageNet64 is a compact dataset designed for Continual Few-Shot Learning (CFSL), developed by the School of Informatics, University of Edinburgh. It contains 200,000 64×64 RGB images spanning 1,000 categories, with 200 images per category. By retaining all categories from the original ImageNet dataset while reducing the number of samples per category, SlimageNet64 aims to provide an efficient and challenging benchmark for CFSL. This dataset is suitable for research on sample efficiency, catastrophic forgetting, continual adaptation of neural networks, and the continual learning capabilities of meta-learning systems, and is particularly ideal for experiments in resource-constrained environments.