Skip to content

Latest commit

 

History

History
55 lines (44 loc) · 2.78 KB

README.rst

File metadata and controls

55 lines (44 loc) · 2.78 KB
Raw

"Guided Image Generation with Conditional Invertible Neural Networks" (2019)

Paper: https://arxiv.org/abs/1907.02392 Supplement: https://drive.google.com/file/d/1_OoiIGhLeVJGaZFeBt0OWOq8ZCtiI7li

Contents

Each subdirectory has its own README file containing the details.

  • experiments/mnist_minimal_example contains code to produce class-conditional MNIST samples in <150 lines total
  • experiments/colorization_minimal_example contains code to colorize LSUN bedrooms in <200 lines total
  • experiments/colorization_cINN contains the full research code used to produce all colorization figures in the paper
  • experiments/mnist_cINN contains the full research code used to produce all mnist figures in the paper

Dependencies

Except for pytorch, any fairly recent version will probably work, these are just the confirmed ones:

Package Version
Pytorch 1.0.X
tqdm >= 4.39.0
Numpy >= 1.15.0
FrEIA >= 0.2.0
Optionally for the experiments:
Matplotlib 2.2.3
Visdom 0.1.8.5
Torchvision 0.2.1
scikit-learn 0.20.3
scikit-image 0.14.2
Pillow 6.0.0

** IMPORTANT NOTE (by morioka) **

mnist_minimal_sample``and ``mnist_cINN may not work on the current FrEIA. They work on [Feb. 20, 2020 version](https://github.com/vislearn/FrEIA/tree/ec47c412ae3cd25277ee1019de8247ad4b4c5db2).

FrEIA sometimes break its backward compatiblity without any notification via version number.