Junyu Gao

remove the "module."

https://github.com/gjy3035/C-3-Framework/issues/22#issuecomment-494287652

This test code is part of our other unpublished paper in fact. I think our test code is only for visualization, so I post it. The model is on ShangHai...

@FanShuixing You would adopt our provided data for training, which is scaled down keep the original ratio. Also, you may need a smaller batch size for single GPU.

Sorry. GCC's training may take much time and many machines. Thus, we have no plan to do these experiments.

可否告知报错内容？谢谢！

1. You may use VPN to download the dataset. 2. This repo provides a potential method of processing GCC. https://github.com/gjy3035/C-3-Framework

由于机器有限，暂时不会提供。 使用__C.PRE_GCC，只会加载模型参数。 使用 __C.RESUME，除了模型参数，训练阶段的任何数据都会加载。这个主要用于由于不可抗力原因导致模型中途训练终止的情况。

Pytorch1.0 is OK. If you want to implement this code in Python 3.6, you should check the "/" operation in our code. SFCN based on VGG-16: [sfcn.zip](https://github.com/gjy3035/GCC-SFCN/files/3101801/sfcn.zip)

@liwenxi Scene 4 is harder than other scenes. Thus, the performance on it is poorer than that of others. I have two suggestions that may improve your performance on it....