Park T, Liu M Y, Wang T C, et al. Semantic image synthesis with spatially-adaptive normalization[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 2337-2346.
Karras T, Laine S, Aila T. A style-based generator architecture for generative adversarial networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 4401-4410.
Wang T C, Liu M Y, Zhu J Y, et al. High-resolution image synthesis and semantic manipulation with conditional gans[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 8798-8807.
Jittor 风景图片生成比赛
简介
本项目包含了第二届计图挑战赛计图 - 风景图像生成比赛的代码实现。本项目基于 SPADE (GauGAN) 实现,主体框架为 VAE-GAN,主要分成 Encoder, Generator, Discriminator 三个部分,训练阶段提供真实图片用于 Encoder 编码,推理阶段采用随机高斯分布采样直接将隐向量输入 Generator。
安装
本项目可在 单张 RTX 3090 上运行,训练时间约为 50 小时。
运行环境
安装依赖
执行以下命令安装 python 依赖
预训练模型
下载地址
训练
单卡训练可运行以下命令:
请调整文件夹格式为
务必保证训练集的 imgs 和 labels 里的文件名一一对应。
推理
生成测试集上的结果可以运行以下命令:
致谢
Park T, Liu M Y, Wang T C, et al. Semantic image synthesis with spatially-adaptive normalization[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 2337-2346.
Karras T, Laine S, Aila T. A style-based generator architecture for generative adversarial networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 4401-4410.
Wang T C, Liu M Y, Zhu J Y, et al. High-resolution image synthesis and semantic manipulation with conditional gans[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 8798-8807.