gan-qp-master
说明： 对于GAN网络的改进，解决了原生GAN的不稳定及会产生模式崩溃的缺点，可以有效地进行图片的生成
(The improvement of the GAN network solves the disadvantages of the instability of the native GAN and the occurrence of mode collapse, which can effectively generate the image.)
(The improvement of the GAN network solves the disadvantages of the instability of the native GAN and the occurrence of mode collapse, which can effectively generate the image.)
文件列表:
128 (0, 2018-12-17)
128\gan_gp_l1.py (8860, 2018-12-17)
128\gan_qp_l2.py (8871, 2018-12-17)
256 (0, 2018-12-17)
256\gan_qp_l1.py (8966, 2018-12-17)
512 (0, 2018-12-17)
512\gan_qp_l1.py (8879, 2018-12-17)
bigan-qp (0, 2018-12-17)
bigan-qp\ae.jpg (617872, 2018-12-17)
bigan-qp\bigan-qp-256.py (10556, 2018-12-17)
bigan-qp\bigan_qp_128.py (7841, 2018-12-17)
128\gan_gp_l1.py (8860, 2018-12-17)
128\gan_qp_l2.py (8871, 2018-12-17)
256 (0, 2018-12-17)
256\gan_qp_l1.py (8966, 2018-12-17)
512 (0, 2018-12-17)
512\gan_qp_l1.py (8879, 2018-12-17)
bigan-qp (0, 2018-12-17)
bigan-qp\ae.jpg (617872, 2018-12-17)
bigan-qp\bigan-qp-256.py (10556, 2018-12-17)
bigan-qp\bigan_qp_128.py (7841, 2018-12-17)
# GAN-QP
GAN-QP: A Novel GAN Framework without Gradient Vanishing and Lipschitz Constraint
arxiv: https://arxiv.org/abs/1811.07296
blog: https://kexue.fm/archives/6163
## environment
The test environment is
- Python 2.7.12
- Keras 2.2.4
- tensorflow 1.8
## 512px samples:
## 256px reconstruction samples (bigan-qp):
## other platforms
- pytorch version: https://github.com/rahulbhalley/gan-qp.pytorch
## 256px reconstruction samples (bigan-qp):
## other platforms
- pytorch version: https://github.com/rahulbhalley/gan-qp.pytorch
近期下载者：
相关文件：
收藏者：