Skip to content
Implementation of "Normalized Diversification (NDiv)" in Pytorch and Tensorflow.
Python
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.
5x5_gaussian
compare_gan @ ac821a9
docs
.gitmodules
README.md
ndiv_pytorch.py
ndiv_tensorflow.py
requirements.txt

README.md

NDiv

Created by Shaohui Liu, Xiao Zhang, Jianqiao Wangni, and Jianbo Shi. This repo includes the source code implementation for "Normalized Diversification" in both PyTorch and TensorFlow.

Introduction

The Normalized Diversification loss, computed over normalized pairwise distance matrices of both the latent vectors and the corresponding outputs in a batchwise manner, was initially described in arXiv to ameliorate the mode collapse problem. We release code on synthetic examples and image generation in this repo. Please follow the instructions in each folder.

prediction example

Citation

@inproceedings{liu2019normalized,
  title={Normalized Diversification},
  author={Liu, Shaohui and Zhang, Xiao and Wangni, Jianqiao and Shi, Jianbo},
  booktitle={Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2019}
}
You can’t perform that action at this time.