https://github.com/pandede/cartoongan
Implement GAN to generate cartoon image
https://github.com/pandede/cartoongan
Last synced: 3 months ago
JSON representation
Implement GAN to generate cartoon image
- Host: GitHub
- URL: https://github.com/pandede/cartoongan
- Owner: Pandede
- Created: 2019-07-29T03:18:28.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2020-10-14T07:16:16.000Z (over 5 years ago)
- Last Synced: 2025-03-22T22:42:50.372Z (over 1 year ago)
- Language: Python
- Size: 390 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
Awesome Lists containing this project
README
# CartoonGAN
Implement GAN to generate cartoon image
## Preparation
1. Download the image data from [Google Drive](https://drive.google.com/file/d/1ohoXETVeFg-LGkzGupwIprKON57dT3Hc/view?usp=sharing) (447MB). The ZIP file includes 10000 500x500 PNG images, as illustrated below:
2. Adjust the required paramters in `config.ini`
```ini
[default]
# Number of training epochs
epoch = 1000
# Save the model each n epochs
save_per_epoch = 20
# Sampling size of generating image each epoch
sample_size = 8
# The image size of generating image
img_size = 64
# The number of image channels (should be always RGB)
img_channel = 3
# The dimension of input noise of generator
noise_size = 32
# Batch size
batch_size = 256
[path]
# The folder path of images
data_src = ./Data/cartoon/
# The folder path of saved models
model_src = ./Pickle/
# The folder path for saving the generated samples
sample_src = ./Sample/
[cuda]
# The device string (training on CPU if device=cpu, or GPU if device=cuda)
device = cuda:0
```
**WARNING**: The CNN layers of generator and discriminator must be modified if `img_size` is changed! The architecture of these model can be modified in `model.py`.
## Run
For PyTorch user`python main_torch.py`
For Keras user `python main_keras.py`
**NOTE** The keras version has not been modularized yet, it may be completed in next several commits.