Filtered Training Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/Blaxzter/Utilizing-Generative-Adversarial-Networks-for-Stable-Structure-Generation-in-Angry-Birds
下载链接
链接失效反馈官方服务：
资源简介：
该数据集经过筛选，包含了用于训练生成对抗网络（GAN）模型的XML层级描述，目的是为游戏“愤怒的小鸟”生成稳定的结构。此外，该数据集的创建旨在确保训练过程中的多样性，并减少GAN训练中的模式崩溃问题。经过过滤后，数据集规模为3566个结构，任务是用GAN生成“愤怒的小鸟”关卡的结构。
This dataset, following screening, contains XML hierarchical descriptions used for training Generative Adversarial Network (GAN) models, with the goal of generating stable structures for the game Angry Birds. Furthermore, this dataset was developed to ensure training diversity and mitigate mode collapse issues during GAN training. After filtering, the dataset consists of 3566 structures, and the assigned task is to generate level structures for Angry Birds using GANs.
提供机构：
Iratus Aves
