Ming

The project was written with tensorflow 1.x. It seems like you are probably running it with tensorflow 2?

请问具体是遇到了哪些问题？

我看一开始你是在和背景噪音相关的issue下回复的。如果你有一些能用于增广的噪音文件的话，可以参考 train901.py，修改31、33行中 data/background 为你自己的路径。如果没有噪音文件只是想训练一下看看效果的话，可以将train901.py中的AUGMENTATION改为False，去掉lib/tools_batch.py中和background相关的代码。

Thanks. It's actually not me who decided to use FC layers, but the authors (https://distill.pub/2020/growing-ca/) chose to use FC. In my opinion, since the outputs of the Sobel filters are...

谢谢支持