GCondenser

GCondenser是一个由昆士兰大学创建的大规模图压缩基准数据集，旨在通过标准化图压缩范式来评估和比较不同的图压缩方法。该数据集包含六个不同的图数据集，涵盖了从小型引文网络到大型图数据集的各种应用场景。数据集的创建过程涉及图的压缩、验证和评估，确保了数据集的质量和实用性。GCondenser的应用领域广泛，包括模型训练效率的提升、跨架构可迁移性的评估以及持续图学习的性能优化。

GCondenser is a large-scale graph compression benchmark dataset developed by The University of Queensland, designed to evaluate and compare diverse graph compression methods through standardized graph compression paradigms. This dataset consists of six distinct graph datasets, covering a broad spectrum of application scenarios ranging from small-scale citation networks to large-scale graph datasets. The construction of this dataset involves graph compression, validation and evaluation procedures, which guarantees its quality and practical applicability. GCondenser has wide-ranging applications, including enhancing model training efficiency, evaluating cross-architecture transferability, and optimizing the performance of continual graph learning.