Unofficial implementation of AttentionGAN with pytorch
Official implementation: https://github.com/Ha0Tang/AttentionGAN
pytorch
torchvision
numpy
openCV2
matplotlib
https://people.eecs.berkeley.edu/~taesung_park/CycleGAN/datasets/
Please refer train.ipynb
Please refer test.ipynb
| Parameter | Value |
|---|---|
| Learning rate | 2e-4 |
| Batch size | 1 |
| Epoch | 100 |
| Optimizer | Adam |
| (beta1, beta2) | (0.5, 0.999) |
| (gamma, lambda1, lambda2) | (10, 10, 1) |
| Data Augmentation | RandomHorizontalFlip() |
| HW | CPU : Intel i5-8400 RAM : 16G GPU : NVIDIA GTX1060 6G |
| Training Time | About 20 hours for 100 epoch |
Please refer the original paper (https://arxiv.org/abs/1903.12296)
