I-RAVEN-Mesh

该数据集通过引入线基模式的全新组成结构，丰富了RPMs，从而在迁移学习环境中更有利于评估知识的逐步获取。此外，该数据集还用于检验模型在部分数据集上的迁移学习能力。其规模根据训练集的比例而变化（例如，q的取值范围在1/64到1之间）。所涉及的任务是评估迁移学习中的逐步知识获取能力。

This dataset enriches RPMs by introducing a novel compositional structure of line-based patterns, thereby facilitating more effective evaluation of gradual knowledge acquisition in transfer learning environments. Furthermore, this dataset is used to examine the transfer learning capabilities of models on partial datasets. Its scale varies according to the proportion of the training set; for example, the value of q ranges from 1/64 to 1. The core task involved in this dataset is to evaluate the ability of gradual knowledge acquisition in transfer learning.