Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Image Inpainting via Generative Multi-column Convolutional Neural Networks, NeurIPS2018
| Date | Stars |
|---|---|
| 2026-07-24 | 435 |
| 2026-07-25 | 435 |
| 2026-07-28 | 435 |
| 2026-07-30 | 435 |
| 2026-07-31 | 435 |
| 2026-08-06 | 435 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Image Inpainting via Generative Multi-column Convolutional Neural Networks
by [Yi Wang](https://shepnerd.github.io/), [Xin Tao](http://www.xtao.website), [Xiaojuan Qi](https://xjqi.github.io), [Xiaoyong Shen](http://xiaoyongshen.me/), [Jiaya Jia](http://www.cse.cuhk.edu.hk/leojia/).
## Results on Places2, CelebA-HQ, and Paris streetview with rectangle masks.
<img src="./pics/teaser.png" width="100%" alt="Teaser">
## Results on Places2 and CelebA-HQ with random strokes.
<p><img src="./pics/places2hd_random.gif" width="65%" alt="places2hd" style="margin-left: auto; margin-right: auto; display: block;"></p>
<p><img src="./pics/celeba-hq.png" width="100%" alt="celeba-hq_512"></p>
## Introduction
This repository is for the NeurIPS 2018 paper, '[Image Inpainting via Generative Multi-column Convolutional Neural Networks](http://papers.nips.cc/paper/7316-image-inpainting-via-generative-multi-column-convolutional-neural-networks.pdf)'.
If our method is useful for your research, please consider citing:
@inproceedings{wang2018image,
title={Image Inpainting via Generative Multi-column Convolutional Neural Networks},
author={Wang, Yi and Tao, Xin and Qi, Xiaojuan and Shen, Xiaoyong and Jia, Jiaya},
booktitle={Advances in Neural Information Processing Systems},
pages={331--340},
year={2018}
}
## Our framework
<img src="./pics/net_architecture.png" width="100%" alt="framework">
### Partial Results
<p><img src="./pics/face1.png" width="48%" alt="face1"> <img src="./pics/face2.png" width="48%" alt="face2"></p>
<p><img src="./pics/face3.png" width="48%" alt="face3"> <img src="./pics/face4.png" width="48%" alt="face4"></p>
<p><img src="./pics/p2.png" width="100%" alt="celeba-hq_512"></p>
<p><img src="./pics/p3.png" width="100%" alt="celeba-hq_512"></p>
### [More results](https://drive.google.com/file/d/1uzYgpiCkENTqW-Yy-AriHXH0wUidi3KY/view?usp=sharing)
## Prerequisites
- Python3.5 (or higher)
- Tensorflow 1.4 (or later versions, excluding 2.x) with NVIDIA GPU or CPU
- OpenCV
- numpy
- scipy
- easydict
- Pytorch 1.0 with NVIDIA GPU or CPU
- tensorboardX
## Installation
```bash
git clone https://github.com/shepnerd/inpainting_gmcnn.git
cd inpainting_gmcnn/tensorflow
```
or
```bash
cd inpainting_gmcnn/pytorch
```
## For tensorflow implementations
### Testing
Download pretrained models through the following links ([paris_streetview](https://drive.google.com/file/d/1wgesxSUfKGyPwGQMw6IXZ9GLeZ7YNQxu/view?usp=sharing), [CelebA-HQ_256](https://drive.google.com/file/d/1zvMMzMCXNxzbYJ_6SEwt3hUShD3Xnz9W/view?usp=sharing), [CelebA-HQ_512](https://drive.google.com/file/d/1cp5e8XyXmHNZWj_piHH4eg4HFi3ICl0l/view?usp=sharing), [Places2](https://drive.google.com/file/d/1aakVS0CPML_Qg-PuXGE1Xaql96hNEKOU/view?usp=sharing)), and unzip and put them into `checkpoints/`. To test images in a folder, you can specify the folder address by the opinion `--dataset_path`, and set the pretrained model path by `--load_model_dir` when calling `test.py`.
For example:
```bash
python test.py --dataset paris_streetview --data_file ./imgs/paris-streetview_256x256/ --load_model_dir ./checkpoints/paris-streetview_256x256_rect --random_mask 0
```
or
```bash
sh ./script/test.sh
```
### Training
For a given dataset, the training is formed of two stages. We pretrain the whole network with only confidence-driven reconstruction loss first, and finetune this network using adversarial and ID-MRF loss along with the reconstruction loss after the previous phase converges.
To pretrain the network,
```shell
python train.py --dataset [DATASET_NAME] --data_file [DATASET_TRAININGFILE] --gpu_ids [NUM] --pretrain_network 1 --batch_size 16
```
where `[DATASET_TRAININGFILE]` indicates a file storing the full paths of the training images.
Then finetune the network,
```shell
python train.py --dataset [DATASET_NAME] --data_file [DATASET_TRAININGFILE] --gpu_ids [NUM] --pretrain_network 0 --load_mExcerpt of 6,205 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:f863a6cabb7f945c, topic:deep-learning, topic:pytorch, topic:tensorflow
matched fp:f863a6cabb7f945c, topic:image-generation, topic:gan, name:inpainting