Skip to content
PyTorch implementation of paper "Flat Metric Minimization with Applications in Generative Modeling"
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
data
figures
results
LICENSE
README.md
datasets.py
demo_2d.py
demo_mnist.py
flatgan.py
models.py
util.py

README.md

Flat Metric Minimization

This repo contains a minimal PyTorch implementation to reproduce Fig. 6 and Fig. 7 from the paper:

Flat Metric Minimization with Applications in Generative Modeling (Thomas Möllenhoff, Daniel Cremers; ICML 2019). https://arxiv.org/abs/1905.04730

Notes

  • We have tested the code on: Ubuntu 16.04; Python 3.7.1; PyTorch 1.0.0
  • Running the MNIST example (demo_mnist.py) will first download the MNIST dataset into the data/ folder
  • The results will be saved in results/2d (for demo_2d.py) and results/mnist (for demo_mnist.py)

Demo outputs

python demo_2d.py --k 0

python demo_2d.py --k 1

python demo_mnist.py 

Publication

@article{flatgan,
    title = {Flat Metric Minimization with Applications in Generative Modeling},
    author={Thomas Möllenhoff, Daniel Cremers},
    journal={International Conference on Machine Learning},
    year={2019},
    url={https://arxiv.org/abs/1905.04730}
}
You can’t perform that action at this time.