chenxy99/SGAN

该数据集专注于视觉-语言问题的解决，采用了一种新颖的Solution Graph Attention Network (SGAN)方法，该方法通过整合步骤内和步骤间的注意力机制，逐步构建解决方案。数据集适用于文本生成任务，主要使用英语，标签包括问题解决，数据大小在10M到100M之间。This dataset focuses on solving vision-language problems using a novel Solution Graph Attention Network (SGAN) approach, which progressively constructs solutions by integrating intra-step and inter-step attention mechanisms. The dataset is suitable for text-generation tasks, primarily in English, with tags including problem solving, and the size category is between 10M and 100M.