# Progressive GAN

Progressive Growing of GANs for Improved Quality, Stability, and Variation<br><i>– Official TensorFlow implementation of the ICLR 2018 paper</i>

### Start new training

- edit `config.py` dataset path
- `desc += '-my_name'; dataset = EasyDict(tfrecord_dir='tf_dir_name'); train.mirror_augment = True;` 
`python train.py`
-

### Resume Training

- `desc += '-my_name'; dataset = EasyDict(tfrecord_dir='512x512'); train.mirror_augment = True; train.resume_kimg = 5587.8; train.resume_run_id = 11;`
- to get `kimg`, `cat 011_yourproject/log.txt` and find latest kimg number