Conditional GAN
Conditional Generative Adversarial Networks for anime generation (AnimeGAN).
Training results dump every 500 min-batch in 25 epoch(26000th min-batch) for the following tags
- blue hair blue eyes
- gray hair green eyes
- green hair red eyes
- orange hair brown eyes
- blonde hair gray eyes
- pink hair aqua eyes
Sample training data
Environment
python3
tensorflow 1.0
scipy
Model structure
Data
Usage
- Download hw3 data from data link, place the MLDS_HW3_dataset/ in the same directory and unzip the face.zip in MLDS_HW3_dataset/
- Replace the tags in MLDS_HW3_dataset/sample_testing_text.txt to the right format.
- Start training !
Train
First time use, you need to do the preprocessing
$ python3 main.py --prepro 1
If you already have done the preprocessing
$ python3 main.py --prepro 0
Model
- dcgan structure
- use one hot encoding for condition tags
Test
This code will automatically dump the results for the tags specified in MLDS_HW3_dataset/sample_testing_text.txt every dump_every batches to the test_img/ folder.
Testing tags format
1,<Color> hair <Color> eyes
2,<Color> hair <Color> eyes
3,<Color> hair <Color> eyes
4,<Color> hair <Color> eyes
.
.
.
- Possible colors for eyes
['<UNK>', 'yellow', 'gray', 'blue', 'brown', 'red', 'green', 'purple', 'orange',
'black', 'aqua', 'pink', 'bicolored']
- Possible colors for hair
['<UNK>', 'gray', 'blue', 'brown', 'red', 'blonde', 'green', 'purple', 'orange',
'black', 'aqua', 'pink', 'white']