yikang-li / pastegan Goto Github PK

OSError: Unable to open file (unable to open file: name = 'data/visual_genome/train.h5', errno = 2, error message = 'No such file or directory', flags = 0, o_flags = 0)

There are some missing files in your folder, which are vocab.json and train.h5. May I know when would you able to upload them?

Hello, may I ask if it is convenient for you to provide your pretrained weight of your full model on VG dataset? Thank you very much.

Hi,

Can you add a license to it? It will be useful.

Thanks,
Krishna

Hi,

Thanks for sharing the code of this great work! Can you provide the source code for preprocessing the data (both vg and coco)? Thanks!

Can you tell me where the semi-parametric and Paste？Thank you！

Hi, can you tell me the tensorflow version? Thank you.

I cannot find the voca.json in the public website of genome and coco , is this file produced in the pre-processing?

Could you kindly provide it for me?

Hello, when I run the train the model as you provided, sometimes it just interrupts beacause of getting an infinite grid, which corresponds to Line 30 in ./utils/layout/boxes_to_layouts.py. Do you know how to cope with this problem? Thank you in advance！

Hi,

Thanks for sharing the code. May I ask what's the purpose of the double perceptual loss here?

It seems not introduced in your paper. Thanks!