Skip to content

reedscot/nips2016

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

###Learning What and Where to Draw Scott Reed, Zeynep Akata, Santosh Mohan, Samuel Tenka, Bernt Schiele, Honglak Lee

This is the code for our NIPS 2016 paper on text- and location-controllable image synthesis using conditional GANs. Much of the code is adapted from reedscot/icml2016 and dcgan.torch.

####Setup Instructions

You will need to install Torch, CuDNN, stnbhwd and the display package.

####How to train a text to image model:

  1. Download the data including captions, location annotations and pretrained models.
  2. Download the birds and humans image data.
  3. Modify the CONFIG file to point to your data.
  4. Run one of the training scripts, e.g. ./scripts/train_cub_keypoints.sh

####How to generate samples:

  • ./scripts/run_all_demos.sh.
  • html files will be generated with results like the following:

Moving the bird's position via bounding box:

Moving the bird's position via keypoints:

Birds text to image with ground-truth keypoints:

Birds text to image with generated keypoints:

Humans text to image with ground-truth keypoints:

Humans text to image with generated keypoints:

####Citation

If you find this useful, please cite our work as follows:

@inproceedings{reed2016learning,
  title={Learning What and Where to Draw},
  author={Scott Reed and Zeynep Akata and Santosh Mohan and Samuel Tenka and Bernt Schiele and Honglak Lee},
  booktitle={Advances in Neural Information Processing Systems},
  year={2016}
}

Releases

No releases published

Packages

No packages published