Yuheng-Li comments

Results 13 comments of


                                            Yuheng-Li

Result is very bad on custom dataset

Hi Ayush, First, what is the second image here? I don't think there are final results. Could you show all fake images? Also, You should set SUPER_CATEGORIES and FINE_GRAINED_CATEGORIES much...

Result is very bad on custom dataset

You should not set super categories or fine-grained categories to be 1. The disentanglement will not learn.

How to print EG_Loss?

You can use tensorboardX to visualize all losses. You can add a few lines of codes around L318 to store eg_loss

What is the Bounding box？

1, bounding boxes are used in the background stage such that the background discriminator can only see background region. 2, If you dataset does not come with box annotations, then...

What is the Bounding box？

Please follow the bounding box format in the CUB dataset http://www.vision.caltech.edu/visipedia/CUB-200-2011.html

Unable to download pretrained models

Hi, We did not move the models, and could you try again?

image resize

Hi, The code is based upon FineGAN ( https://github.com/kkanshul/finegan ) , and this ratio is used there. I did not try the other resolutions, but I recommend use the same...

Unable to setting up the dataset

Hi, please download cub data using link provided by us, and see its format. Like the cub, for your own dataset, you need to prepare two files: images.txt and bounding_boxes.txt...

Checkpoint trained on only 256x256 data?

I have the same request. Thanks

When are you willing to upload the code?

We will release the code in a couple of days. Thanks.