semi-supervised-GAN

semi-supervised-GAN copied to clipboard

请问如何加载本地图片数据集？如CelebA，CK+等

iamzhangyunlu

实现的不对吧，labeled样本并非固定的，每轮迭代都是随机抽取

1

最后还是训练了全部样本

lwang02

你好，我想问下，为什么我跑出来的准确率只有0.21啊

9

xfydybest

question~

4

Hello， when I run the code，I have some touble in understanding the errors that the program generated. Traceback (most recent call last): File "main.py", line 33, in main() File "main.py",...

CXSunshine

about d_loss

1

hi! the l_loss of my running result is -160635.296875,what's wrong?

floricaaa

为啥loss_d是负值？

1

您好，我使用您的默认参数，开始训练为啥loss_d是负值？谢谢 [1.490094][epoch: 0/50][iter: 0/1000],loss_d:-17422.916016,loss_g:1935.270752, d1:17440.179688, d2:0.152927 flag: 1

ljdream0710

你好，我下载的代码中没有deconv2d，conv2d，fc部分的代码，可否发我一份，非常感谢！！！

1

你好，我下载的代码中没有deconv2d，conv2d，fc部分的代码，可否发我一份，非常感谢！！！

floricaaa

zychen2016

More than 200 labeled images are used to train

1

Hi, Thanks for your work. I think there is something wrong in your result.In general, we want to use only 200 labeled images in training process.But in your code, you...

yongpi-scu

train.py里的两个问题（discriminator的loss & 更新频率）

2

### 关于loss https://github.com/LDOUBLEV/semi-supervised-GAN/blob/52ba4a5ebb5ff07e6eceb48897ca78cb6b14f529/train.py#L69-L70 1. 其中69行的 `d_loss_r - d_loss_f` 为什么是减？ 2. 70行 `-d_loss_r` 为什么会有一个负号？ ``还没跑，看着蛮奇怪的。`` ### 关于D更新k次，G更新1次 https://github.com/LDOUBLEV/semi-supervised-GAN/blob/52ba4a5ebb5ff07e6eceb48897ca78cb6b14f529/train.py#L125-L128 ``代码里貌似没有反映出来吧？``

kemingzeng