finish gan

Author: xyliao

README.md

@@ -60,8 +60,8 @@ Learn Deep Learning with PyTorch
 - Chapter 6: 生成对抗网络
     - [自动编码器](https://github.com/SherlockLiao/code-of-learn-deep-learning-with-pytorch/blob/master/chapter6_GAN/autoencoder.ipynb)
     - [变分自动编码器](https://github.com/SherlockLiao/code-of-learn-deep-learning-with-pytorch/blob/master/chapter6_GAN/vae.ipynb)
-    - 生成对抗网络
-    - 深度卷积对抗网络(DCGANs)生成人脸
+    - [生成对抗网络](https://github.com/SherlockLiao/code-of-learn-deep-learning-with-pytorch/blob/master/chapter6_GAN/gan.ipynb)
+    - 深度卷积对抗网络 (DCGANs) 生成人脸
 
 - Chapter 7: 深度增强学习
     - 深度增强学习的介绍
@@ -91,7 +91,7 @@ Learn Deep Learning with PyTorch
     - char rnn 实现文本生成
     - Image Caption: 实现图片字幕生成
     - seq2seq 实现机器翻译
-    - cnn+rnn+attention 实现文本识别
+    - cnn + rnn + attention 实现文本识别
 
 ## 一些别的资源
 

chapter6_GAN/gan.ipynb

@@ -181,7 +181,8 @@
     "ExecuteTime": {
      "end_time": "2018-01-01T10:41:10.409900Z",
      "start_time": "2018-01-01T10:41:10.059597Z"
-    }
+    },
+    "collapsed": true
    },
    "outputs": [],
    "source": [
@@ -301,10 +302,9 @@
     "        im = Variable(im)\n",
     "        if torch.cuda.is_available():\n",
     "            im = im.cuda()\n",
-    "            \n",
-    "        optimizer.zero_grad()\n",
     "        recon_im, mu, logvar = net(im)\n",
     "        loss = loss_function(recon_im, im, mu, logvar) / im.shape[0] # 将 loss 平均\n",
+    "        optimizer.zero_grad()\n",
     "        loss.backward()\n",
     "        optimizer.step()\n",
     "\n",
@@ -401,7 +401,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.2"
+   "version": "3.6.3"
   }
  },
  "nbformat": 4,