Ma Zhenwei

I wonder if the process of generating realistic GCC images are as follows: 1. use 'scale_height_and_crop' to get 360*360 patches for CycleGAN training as your opt.txt above. 2. use trained...

gt is points. I forget to change this annotation.

I change the visualization tool from Visdom to TensorboardX which seems more great. Besides, update the GT density maps generation code which doesn't do well in the previous version. And...

I think more parameters can boost your performance on the public datasets. However, if you do some application for real scenes, it may be overfitting. I think VGG-based architecture is...

The performance in PartA not change. The size of generated densitymap is 1/8 of the original image both in height and width. The new version applys an upsample layer to...

Both.

you can see Model.py. 

Yeah. And the performance on the dataset is pretty well.

sorry，I didn't conduct experiments on world expo. And only do it on Shanghaitech part a. but flip the image and upsample image can enhance the performance.  the learning rate can...

并不是，大力出奇迹，开玩笑的，一般几百个epoch就行了 ---原始邮件--- 发件人: "guodewen"