admin管理员组

文章数量:1124166

I am currently training the BD-Cycle GAN, a modified version of the Mol-Cycle GAN. I have not modified any of the code, but needed to download the utils folder and environment.yml from the Mol-Cycle GAN repo. When running the train.py file with default parameters, the loss for both discriminator A and B is stuck at 0.0, but the generator loss seems to be normal.

I am stuck to what the cause of the issue is since I am using the official repo for the model from the authors, without modifying any of the code or hyperparameters, but getting this result. The loss for the discriminator remains at 0.0 for every epoch, disrupting the entire training process.

本文标签：

版权声明：本文标题：python - Discriminator loss for CycleGAN stuck at 0.0 - Stack Overflow 内容由网友自发贡献，该文观点仅代表作者本人， 转载请联系作者并注明出处：http://www.betaflare.com/web/1736603065a1945263.html， 本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌抄袭侵权/违法违规的内容，一经查实，本站将立刻删除。