/pytorch-inpainting-with-partial-conv

Unofficial pytorch implementation of 'Image Inpainting for Irregular Holes Using Partial Convolutions' [Liu+, arXiv2018]

Primary LanguagePython

pytorch-inpainting-with-partial-conv

Note that this is an ongoing project and I cannot fully reproduce the results. Suggestions and PRs are welcome!

This is an unofficial pytorch implementation of a paper, Image Inpainting for Irregular Holes Using Partial Convolutions' [Liu+, arXiv2018].

Requirements

  • Python3.6
  • Pytorch 0.4.0
pip install -r requirements.txt

Usage

Preprocesse

Generate masks by following [2] (saved under ./masks by default).

Note that the way of the mask generation is different from the original work [1]

python generate_data.py

Train

CUDA_VISIBLE_DEVICES=$(nvidia-empty) python train.py

Results

Here are some results from the test set after the training of 200,000 iterations. Results

TODO

  • [] Fine-tuning (fix BN params in encoder)
  • [] Check the quality

References