This is unofficial implementation of AnoGAN in pytorch
- Ubuntu 18.04
- Python 3.7
- Pytorch 1.8.1
- Set parameters in train.py
- Set gpu(cuda) option in train.py
$ python3 train.py
- Set checkpoint paths in test.py
- Set gpu(cuda) option in test.py
$ python3 test.py
- The left image is input image, right image is generated image from z
2
is trained. Anomaly Score: 0.0276
is not trained. Anomaly Score: 0.067
[1] paper: https://arxiv.org/abs/1703.05921