pix2pix

• An implement of pix2pix for tensorflow version according to paper named Image-to-Image Translation with Conditional Adversarial Nets

Requirements

• tensorflow 1.3.0

• python 2.7.12

• numpy 1.13.1

• scipy 0.17.0

Usages

download repo

  $ git clone https://github.com/nnuyi/pix2pix
  $ cd pix2pix

download datasets

  # here facades indicates dataset name, you can change to Cityscapes etc, example shows below
  $ bash ./dataset_download.sh facades    

• This is repo mainly focus on facades datasets, more datasets you can download here, or you just need to run dataset_download.sh,then, you can see the dataset named facades is stored in your datasets(named datasets) directory.

training model

  $ python main.py --is_training True --phase train

testing model

  $ python main.py --is_testing True --phase test

Results

sample data from val sets(sampling)

sample result	sample result
left:source middle: groundtruth right:sample	left:source middle: groundtruth right:sample

sample data from test sets(testing)

test result	test result
left:source middle: groundtruth right:sample	left:source middle: groundtruth right:sample
left:source middle: groundtruth right:sample	left:source middle: groundtruth right:sample

Acknowledgements

• dataset_down.sh is obtained from phillipi/pix2pix

Contacts

• Email:computerscienceyyz@163.com