资源简介
使用深度对抗网络完成马和斑马的互换,也可以用作打码,或者去掉马赛克。
代码片段和文件信息
import tensorflow as tf
def lrelu(x leak=0.2 name=“lrelu“ alt_relu_impl=False):
with tf.variable_scope(name):
if alt_relu_impl:
f1 = 0.5 * (1 + leak)
f2 = 0.5 * (1 - leak)
# lrelu = 1/2 * (1 + leak) * x + 1/2 * (1 - leak) * |x|
return f1 * x + f2 * abs(x)
else:
return tf.maximum(x leak*x)
def instance_norm(x):
with tf.variable_scope(“instance_norm“):
epsilon = 1e-5
mean var = tf.nn.moments(x [1 2] keep_dims=True)
scale = tf.get_variable(‘scale‘[x.get_shape()[-1]]
initializer=tf.truncated_normal_initializer(mean=1.0 stddev=0.02))
offset = tf.get_variable(‘offset‘[x.get_shape()[-1]]initializer=tf.constant_initializer(0.0))
out = scale*tf.di
属性 大小 日期 时间 名称
----------- --------- ---------- ----- ----
目录 0 2017-12-07 18:29 CycleGAN-master\
文件 39 2017-12-07 18:29 CycleGAN-master\.gitignore
目录 0 2017-12-07 18:29 CycleGAN-master\.ipynb_checkpoints\
文件 5261 2017-12-07 18:29 CycleGAN-master\.ipynb_checkpoints\Blog for CycleGan-checkpoint.ipynb
文件 10511 2017-12-07 18:29 CycleGAN-master\.ipynb_checkpoints\CycleGAN_notebook-checkpoint.ipynb
文件 2076 2017-12-07 18:29 CycleGAN-master\README.md
目录 0 2017-12-07 18:29 CycleGAN-master\__pycache__\
文件 2213 2017-12-07 18:29 CycleGAN-master\__pycache__\la
文件 812 2017-12-07 18:29 CycleGAN-master\download_datasets.sh
目录 0 2017-12-07 18:29 CycleGAN-master\images\
文件 115630 2017-12-07 18:29 CycleGAN-master\images\Generator.jpg
文件 235718 2017-12-07 18:29 CycleGAN-master\images\Results.jpg
文件 70319 2017-12-07 18:29 CycleGAN-master\images\discriminator.jpg
文件 115099 2017-12-07 18:29 CycleGAN-master\images\distortion.jpg
文件 97526 2017-12-07 18:29 CycleGAN-master\images\model.jpg
文件 2533 2017-12-07 18:29 CycleGAN-master\la
文件 14485 2017-12-07 18:29 CycleGAN-master\main.py
文件 4771 2017-12-07 18:29 CycleGAN-master\model.py
- 上一篇:STM32_ADC多通道源码
- 下一篇:STM32ADC双通道采样
评论
共有 条评论